ObjectsThis study aimed to elucidate the effect of cognitive bias modification on depression.MethodsThis research included 10 randomized studies searching four major databases: PubMed, Embase, PsycINFO, and Cochrane Library, with a total sample size of 467. Moreover, they were examined for quality and possible publication bias.ResultsCognitive bias modification (CBM) had statistically significant results, g = −0.64, 95% CI = [−0.97–0.32]. The interpretation of cognitive bias modification shows the highest effect size, g = −1.45, 95% CI = [−2.05–0.88]. When the training place is located in the laboratory, the training effect is significant, g = −1.11, 95% CI = [−1.62–0.61]. The difference is statistically significant when the training environment was changed to home, g = −0.28, 95% CI = [−0.51–0.05]. CBM has a statistical effect on moderate-to-severe depression, g = −0.70, 95% CI = [−1.04–0.36].ConclusionWe found that CBM had a moderate therapeutic effect on depression, whether the setting was at home or in the lab. Especially when the interpretation of cognitive bias modification (CBM-I) was used, we got the highest effect value. Furthermore, CBM has a statistical effect on moderate-to-severe depression.