Main results We included five RCTs (733 women) comparing exercise with no active treatment, exercise with yoga and exercise with HT. The evidence was of low quality: Limitations in study design were noted, along with inconsistency and imprecision. In the comparison of exercise versus no active treatment (three studies, n = 454 women), no evidence was found of a difference between groups in frequency or intensity of vasomotor symptoms (SMD-0.10, 95% CI-0.33 to 0.13, three RCTs, 454 women, I 2 = 30%, low-quality evidence). Nor was any evidence found of a difference between groups in the frequency or intensity of vasomotor symptoms when exercise was compared with yoga (SMD-0.03, 95% CI-0.45 to 0.38, two studies, n = 279 women, I 2 = 61%, low-quality evidence). It was not possible to include one of the trials in the meta-analyses; this trial compared three groups: exercise plus soy milk, soy milk only and control; results favoured exercise relative to the comparators, but study numbers were small. One trial compared exercise with HT, and the HT group reported significantly fewer flushes in 24 hours than the exercise group (mean difference 5.8, 95% CI 3.17 to 8.43, 14 participants). None of the trials found evidence of a difference between groups with respect to adverse effects, but data were very scanty. Authors' conclusions Evidence was insufficient to show whether exercise is an effective treatment for vasomotor menopausal symptoms. One small study suggested that HT is more effective than exercise. Evidence was insufficient to show the relative effectiveness of exercise when compared with HT or yoga.