How Accurate Is Samsung Galaxy Watch and Whoop Blood Pressure?

Close to a cuff when your pressure is near the readings you calibrated with, and less accurate as it moves away. Galaxy Watch and Whoop both estimate blood pressure from the optical pulse signal and use readings from your own upper-arm cuff to turn it into numbers. Samsung asks for three cuff readings every 28 days, and Whoop for three at setup and one every 30 days after. Samsung's own validation in 85 people measured accuracy 30 minutes after calibration, when the anchor is freshest. Independent studies of Galaxy Watches found good average agreement at rest, readings pulled toward the calibration value, and larger errors once true pressure had moved about 10 points away. Whoop's figures come from its own data and include no test of how its estimates follow a real change. Samsung's instructions also rule out the feature for people with diabetes, an irregular heart rhythm or other heart disease.

What calibration does

An optical sensor reads the shape and timing of each pulse, which changes with the stiffness and tone of the blood vessels. It does not read pressure. Calibration supplies the missing link, pairing your pulse signal with cuff readings taken at the same time so the device can translate later pulse signals into millimeters of mercury. Every reading after that is an estimate of how far your pressure has moved from those calibration values.

On a Galaxy Watch, calibration means three cuff readings within 30 minutes, with the cuff on one arm and the watch on the other wrist, entered into the Samsung Health Monitor app. Samsung's instructions ask for 5 minutes of seated rest first and no caffeine, alcohol, smoking, exercise or bathing in the 30 minutes before. The calibration belongs to one person, and Samsung says a watch calibrated for someone else should not be used. Readings are taken when you press Measure, sitting still with the arm on a table. Whoop asks for three cuff readings at setup and one every 30 days, and gives a morning systolic and diastolic range estimated from overnight heart rate, HRV and pulse patterns.

Samsung also limits the range of the anchor itself. Calibration readings have to fall between 80 and 170 systolic and 50 and 110 diastolic, and the watch reports between 70 and 180 systolic and 40 and 120 diastolic.

What Samsung's own validation tested

Samsung's instructions report one clinical trial of 85 people at a single site between July and October 2019, comparing the watch with cuff readings taken by two medical professionals. The watch averaged 0.11 below the reference for systolic pressure with a spread of 7.46, and 0.28 below for diastolic with a spread of 5.85. The cuff standard allows an average error of up to 5 with a spread of up to 8, and Samsung describes its protocol as a modified version of that standard.

Every reading in that trial was taken 30 minutes after calibration, when the anchor is freshest and pressure has had the least time to move. The result therefore says little about day 20 of a 28-day calibration, or about a morning when pressure is well above where it was. Samsung describes the output as an indication of trends between regular checks by a healthcare professional, for adults 22 and older.

What independent studies of Galaxy Watches found

Over a full 28-day calibration period, 37 people compared a Galaxy Watch 5 with an Omron upper-arm cuff every day. Average agreement was close, within half a point systolic and 0.62 diastolic, with little drift across the month, and the watch met the ISO and IEEE accuracy criteria on all but one measure. Once true pressure was 10 points away from the calibration value, though, the average error rose to 3.4 systolic and 5.1 diastolic, and the authors recommend confirming with a cuff when pressure swings substantially (Digital Health, 2026).

An earlier Galaxy Watch Active 2 worn alongside 24-hour monitoring in 40 people showed the same pull in a stronger form. Low pressures read high and high pressures read low, as if drawn toward the calibration point. As a test for hypertension it caught 83 percent of cases but correctly cleared only 41 percent of people without it, and the authors judged it not ready for clinical use (Frontiers in Cardiovascular Medicine, 2022).

In 50 adults aged 50 to 89, averaging 71, a Galaxy Watch 6 agreed well with an ambulatory cuff across three resting readings, with intraclass correlations of 0.88 for both systolic and diastolic pressure (JMIR Formative Research, 2026). It is one of the few tests in older adults, and the authors limit it to controlled resting conditions in people without decompensated illness. The authors of all three studies declare no conflicts of interest.

Why accuracy falls as pressure moves from the anchor

The same pattern turns up in other calibrated devices, which points to the calibration method itself. In 166 people wearing a pulse-transit cuffless device and a 24-hour cuff monitor at the same time, the cuff's 24-hour average sat 7.4 points systolic away from the calibration reading on average, while the cuffless device's sat only 1.8 away. The further pressure had actually moved, the larger the gap between the two devices (Journal of Hypertension, 2024).

The Hilo band, calibrated with a cuff, shows what that means after a medication change. In three patients whose blood pressure medication was increased, the home cuff recorded an average fall of 19.7 systolic and 11.5 diastolic over two weeks, while the band recorded 1.0 and 0.8 (Journal of Hypertension, 2023). The sample is tiny, and it is also the situation in which a reader most needs the number to move.

Some optical devices do follow change. A research PPG wristband from the Swiss engineering center CSEM followed medication-driven changes in 30 patients over six weeks, matching the direction of large changes 83 to 86 percent of the time without recalibration (Blood Pressure, 2026). It is not a consumer product. The FDA's January 2026 draft guidance for cuffless devices now asks for exactly this kind of change test, with pressure at least 15 points systolic or 10 diastolic from the calibration value and at least 30 percent of measurements in each direction. Neither Samsung nor Whoop has published one.

What Whoop has published

Whoop's accuracy page reports 10,746 paired measurements against an FDA-cleared cuff, with an average absolute error of 5.03 systolic and 3.64 diastolic, and 88.68 percent of systolic and 96.57 percent of diastolic estimates within 10 points. It does not give the number of people, the spread of the errors, or how estimates held up when pressure moved away from the calibration readings.

The 5.03 figure is easy to misread against the cuff standard. The standard's limit of 5 applies to the average error, where misses above and below cancel out, and pairs it with a limit on the spread. An average absolute error counts every miss as positive, so it measures something different, and the two numbers cannot be compared directly.

The FDA sent Whoop a warning letter over Blood Pressure Insights on July 14, 2025, on the grounds that a blood pressure estimate is inherently tied to diagnosing high and low blood pressure and the feature had no FDA clearance. On June 17, 2026, after Whoop changed the feature and its labeling, the FDA closed the matter under its January 2026 general wellness guidance and said it would not enforce device requirements on the modified product. The letter applies to that modified feature only and is not a clearance.

Who Samsung says should not use it

Samsung's instructions for use list a long set of contraindications. The app is not for people under 22 or who are pregnant, or for anyone with an arrhythmia, prior heart failure or heart attack, peripheral vascular disease or poor circulation, valve disease, cardiomyopathy, other known cardiovascular disease, end-stage kidney disease, diabetes, a neurological disorder such as tremor, a clotting disorder or prescribed blood thinners, or a tattoo on the watch wrist. The same document says not to change medication or dosage based on the readings without a doctor.

Several of those conditions, including diabetes, atrial fibrillation and use of blood thinners, become more common with age. The readers most likely to want a blood pressure number are often the ones Samsung's instructions exclude.

What this does not settle

No published study tests a Galaxy Watch or Whoop through a large change in pressure the way the FDA draft and the European Society of Hypertension recommend, and none follows readers over several calibration cycles. Whoop's figures have not been published in a journal. Results on one Galaxy Watch generation do not automatically carry over to the next.

Common questions

KnowYourPrime Evidence Registry: Finding KYP-ER-0022: Do Calibrated Cuffless Blood Pressure Readings Track Real Changes? · Finding KYP-ER-0001: Smartwatch Blood Oxygen (SpO2) Accuracy Against Arterial Blood (SaO2) · Finding KYP-ER-0021: Cuffless Blood Pressure Accuracy at Night