Apple Watch ovulation tracking: how accurate is it?

The short answer

• 80%

  Of estimates made during the cycle were within 2 days of ovulation, when there was a clear temperature rise

• 1.2days

  Average error once the cycle had ended, in those cycles

• 4days

  Typical wait after ovulation before an estimate appeared

• 13%

  Of 612,613 real cycles were the “textbook” 28 days long

Key findingWrist temperature rises after ovulation, not before it. So Apple Watch can estimate when you probably ovulated, a few days later. It can’t tell you in advance, it can be wrong, and it isn’t birth control.

How wrist temperature shows ovulation

Ovulation is when an ovary releases an egg, part-way between one period and the next. After it, your body makes more of a hormone called progesterone. That makes your temperature at rest go up a little, by about 0.2 °C or more.

The classic way to see this is basal body temperature (BBT). You take your temperature by mouth every morning, before you get out of bed. A wrist sensor does a similar job while you sleep. Apple Watch Series 8 and later check your wrist every 5 seconds overnight. After about 5 nights, they show each night as a change from your usual.

The catch is timing. The rise comes after ovulation, so it only tells you looking back. In a study of 136 women who wore a wrist sensor in bed, a clear rise showed up in 82% of cycles. In 86% of those cycles, it came on or after the day of ovulation.

Some people look for the lowest point as a sign that ovulation is close. In that study, it fell in the 6 days up to ovulation in only 41% of cycles. The pattern still showed through after alcohol, a big meal, sex, coffee or exercise before bed.

Ovulation isn’t always on day 14

A common belief is that ovulation happens on day 14. Real cycles vary much more. A study of 612,613 cycles from 124,648 users of a cycle-tracking app found that only 13% were exactly 28 days long. The average cycle was 29.3 days.

The first part of the cycle, from your period to ovulation, changes the most. So counting days only gives a rough guess. That is the gap wrist temperature tries to fill.

Apple Watch ovulation tracking: what Apple’s study found

This was a cohort study: a study that followed a group of people over time. Apple paid for it, and half of the 10 authors work for Apple. Ovulation day was set as the day after a positive LH test. LH is a hormone that surges about 1 to 1.5 days before ovulation.

A “clear rise” means a rise of at least 0.2 °C. In those cycles, the watch gave an estimate about 8 times in 10. Counting every cycle, it gave one about 7 times in 10, and the average error during the cycle grew to about 1.9 days.

Next-period predictions were close too. About 88% came within 3 days of the real start, against 84% when counting days alone.

Does it work with irregular cycles?

Mostly, yes, though a little less well. The study split people into two groups. In one, every cycle was 23 to 35 days long. In the other, at least one cycle was shorter or longer. The estimates were slightly less exact for the second group.

The study left out people using hormonal birth control and people who were breastfeeding. It didn’t look separately at conditions such as PCOS (polycystic ovary syndrome) or thyroid problems. Fevers were rare, in 4% of cycles, so we don’t know how much illness throws the estimate off.

Can it be wrong?

Yes, in two ways. First, it can miss: even in cycles with a clear rise, about 1 in 5 had no estimate. Second, it can give an estimate when there was nothing to find. The researchers fed in only random sensor “noise”, with no real rise. The method still gave an estimate in 8% of ongoing cycles and 15% of finished ones.

The authors say the estimate “should not be used in isolation”. Their urine-test method can be wrong at times too. In PCOS, for example, a hormone surge isn’t always followed by ovulation.

Wrist sensor or morning thermometer?

A Swiss study compared the two in 57 women over 193 cycles. Each night they wore a wrist sensor, and each morning they took their temperature by mouth. Urine tests showed which cycles had ovulation.

The wrist sensor found the rise in far more cycles with ovulation. But it also showed a “rise” in 17 of the 23 cycles where urine tests found no ovulation. On average, the first rise came 4.4 days after ovulation on the wrist and 2.7 days after with the thermometer.

Apple’s study found something similar. When fed morning mouth readings instead, Apple’s method gave estimates in fewer cycles than with the wrist (64% and 69%).

One thing is still missing. A 2026 review of Apple Watch accuracy found no studies yet that had checked the watch’s sensor against a reference method.

Claims and evidence

• ClaimApple Watch can tell you your fertile days in advance.

  EvidenceWrist temperature rises after ovulation. Apple’s estimate typically came 4 days later, and in another study the first wrist rise came 4.4 days after.

• ClaimEveryone ovulates on day 14.

  EvidenceOnly 13% of 612,613 cycles were 28 days long. In 31–35-day cycles, ovulation came after about 19.5 days on average.

• ClaimIf your watch shows an ovulation estimate, you definitely ovulated.

  EvidenceApple’s method gave estimates from sensor noise alone in 8% to 15% of cycles. A wrist sensor showed false rises in 17 of 23 cycles without ovulation.

• ClaimA morning thermometer is always more accurate than a wrist sensor.

  EvidenceIn 57 women, the wrist sensor caught the rise in 62% of cycles with ovulation, against 23% for a mouth thermometer. But it also gave more false rises.

What this means for you

What the research can’t tell you yet

• Company links. Apple paid for its study. The wrist-sensor studies were run with Ava, and the 600,000-cycle study by Natural Cycles.
• No scans. Ovulation was judged with home urine tests, not ultrasound scans, so some cycles may be mislabelled.
• Who was studied. People not using hormones. We don’t know how well it works with PCOS or thyroid problems, or while breastfeeding.
• The sensor itself. No study has yet checked Apple Watch wrist temperature against a reference method.
• After the fact. Temperature only shows ovulation once it has happened.

How Bodyvault uses this research

Bodyvault’s Cycle feature estimates your phase from the period dates you log and says plainly that it isn’t a contraceptive. It doesn’t use wrist temperature and never shows fertile windows or “safe days”. It can import periods and ovulation-test results from Apple Health, and suggests seeing a doctor after 90 days without a period. In Bodyvault Pro, your Readiness score compares wrist temperature after ovulation with the same phase of past cycles. That way, the normal rise doesn’t count against you.

General information about fitness and nutrition, not medical advice. If you have a health condition, are pregnant or take medication, talk to your doctor before changing your diet or training.