Can a Smartwatch or Ring Really Measure Blood Pressure?

Only one kind can measure it the way a home monitor does. Watches with a small inflatable cuff in the strap, the Huawei Watch D line and Omron's HeartGuide, have passed the international cuff standard at rest, though independent tests outside the lab found larger errors. Everything else uses an optical pulse sensor, and that comes in two forms. Samsung and Whoop show numbers and the Hilo band shows trends, but only after you calibrate them with a real cuff, and independent studies find their readings stay pulled toward that calibration value, drift at night, and can miss a real change such as a fall after a medication increase. Apple, Oura, Google and RingConn show no numbers at all and look for a pattern over about a month. Apple's version is FDA-cleared, and in its validation study of 1,863 people it caught 41.2 percent of those with hypertension while rarely alerting people without it. The American Heart Association and the European hypertension societies say none of these optical features should be used to diagnose or manage high blood pressure yet.

What a watch or ring is actually measuring

They differ in what they physically measure. An inflatable cuff squeezes the artery and reads the pressure, the same principle as an upper-arm home monitor. An optical sensor reads the shape and timing of the pulse, which is related to blood pressure through the stiffness and tone of the blood vessels but is not pressure itself. To turn that into millimeters of mercury, a device needs either your own cuff readings as an anchor or a general model that has to hold for everyone who wears it.

So the products sort into the groups below, and a number from one group and an alert from another cannot be compared directly.

Why calibrated readings drift away from the truth

A calibrated optical estimate knows your pressure at the moment you calibrated. Between calibrations it has to infer any change from the pulse waveform, and that is where independent studies find the weakness. Samsung asks for cuff calibration every 28 days, and Whoop for three cuff readings at setup and one every 30 days after. Samsung's own validation and the studies of each Galaxy Watch generation are covered in how accurate Samsung and Whoop blood pressure is.

On a Galaxy Watch Active 2 worn alongside 24-hour ambulatory monitoring in 40 people, readings showed a bias toward the calibration point, overestimating low pressures and underestimating high ones (Frontiers in Cardiovascular Medicine, 2022). Over 28 days on a Galaxy Watch 5 in 37 people, average agreement with a cuff was close, but once true pressure was 10 points away from the calibration value the average error rose to 3.4 systolic and 5.1 diastolic, and the authors concluded accuracy declined as values diverged from the calibration point (Digital Health, 2026).

The Hilo band (formerly Aktiia) agreed with 24-hour monitoring during the day but read night-time systolic pressure 15.5 points higher and understated the overnight dip in 41 people. In three patients whose medication was increased, the home cuff recorded an average fall of 19.7 systolic and 11.5 diastolic, and the band recorded 1.0 and 0.8 (Journal of Hypertension, 2023). A pulse-transit device in 166 people showed the same pattern: the further pressure moved from the calibration reading, the larger the gap between it and the cuff (Journal of Hypertension, 2024).

Across studies of wearable cuffless devices against 24-hour monitoring, daytime averages agreed within acceptable limits but night-time readings were off by 4.48 systolic and 5.64 diastolic on average, and the authors concluded the devices are not yet reliable for night-time blood pressure (American Journal of Preventive Cardiology, 2025).

A systematic review of 245 studies of wearable blood pressure, most using pulse transit time or pulse wave velocity, found accuracy moderate to high for diastolic pressure but more variable for systolic, with calibration still needed and external validation rare (Archives of Gynecology and Obstetrics, 2026).

In 67 people wearing a watch-like optical sensor on the arm or wrist alongside a standard 24-hour cuff monitor, 24-hour average systolic readings differed by 1.8 mm Hg on average, but individual differences ran from 13.9 below to 10.3 above (Scientific Reports, 2023).

Watches with a real cuff in the strap

The Huawei Watch D line hides a small air bladder in the strap that inflates around the wrist, so it measures pressure directly and needs no calibration. Omron's HeartGuide, cleared by the FDA in 2018, works the same way. In 85 adults the original Watch D averaged 1.4 systolic and 0.2 diastolic below the reference, with spreads of 6.47 and 5.85, and met the international cuff standard (Blood Pressure Monitoring, 2022). The Watch D2 met it again in 85 adults at rest, averaging 1.6 systolic and 1.9 diastolic below the reference with spreads of 6.4 and 5.0, and agreed with 24-hour ambulatory monitoring in 35 adults, including at night (Hypertension Research, 2025).

The standard allows an average error of up to 5 with a spread of up to 8, so a watch can pass it and still be several points off on any one reading. A wrist reading also depends on where the wrist sits relative to the heart, which the European cuffless protocol tests separately.

Outside those validations the results are less clean. In 41 outpatients wearing a Watch D2 and a validated 24-hour monitor at the same time, the watch read systolic pressure 6.8 points low on average, and only 47 percent of systolic readings landed within 10 points of the reference, below the standard's acceptability threshold (Revista Clínica Española, 2026). In 100 patients from an ordinary clinic, Omron's HeartGuide, another watch with a cuff in the strap, read systolic pressure about 8.5 points low and did not pass all of the standard's criteria (Clinical Research in Cardiology, 2024). The European Society of Hypertension's 2026 consensus statement on these watches judges the static validations generally acceptable and the evidence for 24-hour use limited (Journal of Hypertension, 2026).

A ring cleared without cuff calibration

In September 2026 the FDA cleared Happy Health's blood pressure software for its Happy Ring, which reads light, skin conductance, motion and temperature from a ring on the index finger and reports systolic and diastolic numbers with no cuff calibration and no age or sex input. The clearance covers spot checks in adults aged 22 to 50 at home or in a clinic. It excludes people with arrhythmias such as atrial fibrillation, severe vascular disease or Raynaud's, people taking vasoactive drugs, pregnancy, and readings taken during motion.

The accuracy evidence is the company's own study as summarized by the FDA: 85 people compared against an arterial line, meeting both criteria of the cuff standard. No independent study has appeared in PubMed. A resting comparison also cannot show whether readings follow a change in pressure, the weakness the European cuffless tests were written to catch.

Alerts without numbers: Apple, Oura and Google

Apple Watch Series 9 and later and Ultra 2 and later can send a hypertension notification. Apple says the feature does not measure blood pressure or require calibration; it analyses 60-second optical samples taken about every two hours while you sit still during the day, over 30-day windows, and is for adults 22 and over with no prior diagnosis, not during pregnancy. The FDA cleared it in September 2025. In Apple's study of 1,863 people, it alerted 41.2 percent of those with hypertension, 29.6 percent at stage 1 and 53.7 percent at stage 2, and did not alert 92.3 percent of those without it.

An alert and a silent watch carry very different weight. Take a group where 40 in 100 people have hypertension they do not know about. Running Apple's own figures, about 78 in 100 of those alerted would have it, and about 30 in 100 of those who got no alert would still have it.

Oura added Blood Pressure Signals and Nighttime BP for Gen3 and later rings with a membership. Signals looks at 30 days of nighttime pulse data, needs at least 15 nights, and only notifies when it finds a pattern; Nighttime BP sorts the overnight dip into typical, reduced, pronounced or rising, without numbers, and the Oura blood pressure guide covers what each category has meant in cuff research. Oura's own blog reports 62 percent sensitivity and 91 percent specificity for hypertension, and for dipping 84 percent and 69 percent in 134 people against 48-hour monitoring; these figures have not been published in a journal. Google's blood pressure trends on Pixel Watch 3 and later show High, Not high or Not enough data, and Google states the feature is not a prescreener for hypertension. RingConn Gen 3 shows trends and states it is not a medical device.

What counts as validated, and what the medical societies say

The cuff standard tests whether readings agree on average and in spread. For cuffless devices the European Society of Hypertension recommends up to six tests, including whether readings follow a fall in pressure after treatment, a rise during exercise, the change from awake to asleep, and whether calibration holds over time (Journal of Hypertension, 2023). The FDA's January 2026 draft guidance for cuffless devices asks for a change test in which pressure moves at least 15 points systolic or 10 diastolic away from the calibration value. Whoop's published figures and Samsung's announcement do not report one.

The American Heart Association's scientific statement concludes that the devices on the market have not yet been adequately vetted for accuracy and efficacy (Hypertension, 2026). The European Society of Hypertension notes that its guidelines do not recommend cuffless devices for the diagnosis and management of hypertension (Journal of Hypertension, 2022), and a 2026 European scientific statement does not recommend their use in clinical decisions (European Journal of Preventive Cardiology, 2026).

Why most launched without FDA clearance: the FDA's January 2026 wellness guidance allows an optical blood pressure value as a general wellness output if it is validated and the product makes no diagnostic claims. Samsung, Whoop, Oura, Google and RingConn all label their features as wellness or not for diagnosis. Apple's notification, Happy Health's ring software and Aktiia's G0 system, the clinical version of the Hilo band, went through FDA 510(k) clearance. The Hilo Core sold to US consumers since July 2026 is labeled general wellness, and Hilo states the clearance does not apply to it.

The strongest validation result so far is for a finger device, not a watch. A finger-worn optical monitor met all seven European Society of Hypertension validation scenarios in 768 people, including recalibration, exercise and asleep tests, with mean differences of 5 mm Hg or less (Journal of the American Heart Association, 2026).

What this does not settle

Much of the validation is short, at rest, and often funded by or linked to the manufacturer. Few studies test the same device over months, during sleep, or after a real change such as starting medication, which is when a home reader most wants the number to move. Results for one model do not carry over to the next.

Common questions

KnowYourPrime Evidence Registry: Finding KYP-ER-0020: Accuracy of Smartwatches With a Built-In Blood Pressure Cuff · Finding KYP-ER-0021: Cuffless Blood Pressure Accuracy at Night · Finding KYP-ER-0022: Do Calibrated Cuffless Blood Pressure Readings Track Real Changes?