Several prospective cohort studies have examined the association between prediagnosis and/or postdiagnosis physical activity (PA) on colorectal cancer outcomes and reported conflicting results. To quantitatively assess this association, we have conducted a meta-analysis of prospective studies. Databases and reference lists of relevant studies were searched using MEDLINE and EMBASE up to January 2013. Pooled relative risks (RRs) with 95% confidence intervals (CIs) were calculated using random-effects models. For this meta-analysis, a total of seven prospective cohort studies were included. The analysis included 5,299 patients for prediagnosis PA and 6,348 patients for postdiagnosis PA, followed up over a period ranging from 3.8 to 11.9 years. The analyses showed that patients who participated in any amount of PA before diagnosis had a RR of 0.75 (95% CI: 0.65-0.87, p < 0.001) for colorectal cancer-specific mortality compared to patients who did not participate in any PA. Those who participated in high PA before diagnosis (vs. low PA) had a RR of 0.70 (95% CI: 0.56-0.87, p = 0.002). Similarly, patients who participated in any PA after diagnosis had a RR of 0.74 (95% CI: 0.58-0.95, p = 0.02) for colorectal cancer-specific mortality compared to patients who did not participate in any PA. Those who participated in high PA after diagnosis (vs. low PA) had a RR of 0.65 (95% CI: 0.47-0.92, p = 0.01). Similar inverse associations of prediagnosis or postdiagnosis PA were found for all-cause mortality. In conclusion, both prediagnosis and postdiagnosis PA were associated with reduced colorectal cancer-specific mortality and all-cause mortality. Keywords: cancer; colon; exercise; mortality; physical activity; rectum.