DNA mismatch repair (MMR) status was considered to be a potential prognostic factor for colorectal cancer (CRC) but with conflicting reports, and varied in terms of TNM stages. Its relationship with prognosis in stage II-III CRC had not yet been systematically established. Therefore, we retrieved eligible studies published through May 2019, and screened out 51 studies that reported survival data (overall survival [OS] and/or disease-free survival [DFS]) in 28 331 CRC patients at stage II-III, totally 16.4% of whom were characterized as deficient MMR (dMMR). Significant associations of dMMR status were observed with longer OS (Hazard Ratio [HR] = 0.74, 95% CI: 0.68-0.82; P < .001), as well as DFS (HR = 0.67, 95% CI: 0.59-0.75, P < .001). However, dMMR patients received no statistically significant benefit from fluoropyrimidine-based treatment for either OS (HR = 0.84, 95%CI: 0.60-1.17; P = .31) or DFS (HR = 0.83, 95%CI: 0.60-1.15; P = .27), compared with that in proficient MMR (pMMR) patients for both OS (HR = 0.55, 95% CI: 0.43-0.71; P < .001)and DFS (HR = 0.60, 95% CI: 0.50-0.73; P < .001). Our analysis indicate that dMMR CRC patients at stage II-III had higher OS and DFS than pMMR ones, and fluoropyrimidine-based chemotherapy could improve survival in pMMR patients rather than dMMR ones.chemotherapy, colorectal cancer, DNA mismatch repair, prognosis † Zhujun Deng and Yun Qin contributed equally to this work