Objective: Liver cancer is the second most common cause of death from cancer. Physical activity (PA) was found to be associated with lower risks of several types of cancer. However, the association between PA and the risk of liver cancer is still inconclusive. This systematic review and meta-analysis was aiming to summarize the association between PA and liver cancer risk. Methods: Literatures related were identified by searching PubMed, EMBASE, and Chinese Biomedical literature database from 1965 to 2017 without language limitation. Meta-analyses were performed using random effect model. Results: A total of 5 cohort studies involving 2 513 975 subjects were identified. The pooled relative risk of leisure-time PA with liver cancer risk was 0.92 [95% confidence interval (CI), 0.84-1.01]. There is no significant association between leisure-time PA and liver cancer risk. However, leisure-time PA significantly reduced liver cancer risk in never smokers. The pooled hazard ratio of daily total PA with liver cancer risk was 0.75 (95% CI, 0.66-0.86). Conclusions: Daily total PA significantly reduces liver cancer risk, whereas leisure-time PA significantly reduces liver cancer risk only in never smokers.