Published literature reports controversial results about the association of physical activity (PA) with risk of hypertension. A meta-analysis of prospective cohort studies was performed to investigate the effect of PA on hypertension risk. PubMed and Embase databases were searched to identify all related prospective cohort studies. The Q test and I(2) statistic were used to examine between-study heterogeneity. Fixed or random effects models were selected based on study heterogeneity. A funnel plot and modified Egger linear regression test were used to estimate publication bias. Thirteen prospective cohort studies were identified, including 136,846 persons who were initially free of hypertension, and 15,607 persons developed hypertension during follow-up. The pooled relative risk (RR) of main results from these studies suggests that both high and moderate levels of recreational PA were associated with decreased risk of hypertension (high versus low: RR, 0.81; 95% confidence interval, 0.76-0.85 and moderate versus low: RR, 0.89; 95% confidence interval, 0.85-0.94). The association of high or moderate occupational PA with decreased hypertension risk was not significant (high versus low: RR, 0.93; 95% confidence interval, 0.81-1.08 and moderate versus low: RR, 0.96; 95% confidence interval, 0.87-1.06). No publication bias was observed. The results of this meta-analysis suggested that there was an inverse dose-response association between levels of recreational PA and risk of hypertension, whereas there was no significant association between occupational PA and hypertension.