Several large epidemiological studies have reported an inverse relationship between blood pressure and physical activity. However, longitudinal intervention studies are more appropriate for assessing the effects of physical activity. We performed meta-analyses of randomized controlled trials involving dynamic aerobic endurance training or resistance training. The meta-analysis on endurance training involved 72 trials and 105 study groups. After weighting for the number of trained participants, training induced significant net reductions in resting and daytime ambulatory blood pressure of, respectively, 3.0/2.4 mmHg (P<0.001) and 3.3/3.5 mmHg (P<0.01). The reduction in resting blood pressure was more pronounced in the 30 hypertensive study groups (-6.9/-4.9) than in the others (-1.9/-1.6; P<0.001 for all). Systemic vascular resistance decreased by 7.1% (P<0.05), plasma norepinephrine by 29% (P<0.001), and plasma renin activity by 20% (P<0.05). Body weight decreased by 1.2 kg (P<0.001), waist circumference by 2.8 cm (P<0.001), percentage body fat by 1.4% (P<0.001) and the homeostasis model assessment index of insulin resistance by 0.31 units (P<0.01); high-density lipoprotein cholesterol increased by 0.032 mmol/l (P<0.05). Resistance training has been less well studied. A meta-analysis of nine randomized controlled trials (12 study groups) on mostly dynamic resistance training revealed a weighted net reduction in blood pressure of 3.2 (P=0.10)/3.5 (P<0.01) mmHg associated with exercise. Endurance training decreases blood pressure through a reduction in systemic vascular resistance, in which the sympathetic nervous system and the renin-angiotensin system appear to be involved, and favourably affects concomitant cardiovascular risk factors. The few available data suggest that resistance training can reduce blood pressure. Exercise is a cornerstone therapy for the prevention, treatment and control of hypertension.