Background: Atypical antipsychotic(AAP) augmentation is an alternative strategy for patients with major depressive disorder(MDD) who had an inadequate response to antidepressant therapy. We aimed to compare and rank the efficacy and safety of 4 AAPs in the adjuvant treatment of MDD. Methods: We searched randomized controlled trials(RCTs) published and unpublished from the date of databases and clinical trial websites inception to April 30, 2022. The risk of bias and certainty of the evidence is assessed using the Cochrane bias risk tool and GRADE framework, respectively. Based on the random effects model, we estimated summary risk ratios (RRs) or standardized mean difference (SMD) using network meta-analysis. This study is registered with PROSPERO, number CRD42012002291. Results: 57 eligible RCTs comprising 10900 participants were included. In terms of primary efficacy outcome, compared with placebo, all AAPs had significant efficacy ( SMD= -0.40; 95% CI, -0.68 to -0.12 for quetiapine; -0.35, -0.59 to -0.11 for olanzapine; -0.28, -0.47 to -0.09 for aripiprazole and -0.25, -0.42 to -0.07 for brexpiprazole, respectively). In terms of acceptability, no significant difference was found, either agents versus agents or agents versus placebo. In terms of tolerability, compared with the placebo, quetiapine (RR= 0.24; 95% CI,0.11-0.53), olanzapine (0.30,0.10-0.55), aripiprazole (0.39,0.22-0.69), and brexpiprazole (0.37,0.18-0.75) were significantly less well tolerated. 8 (14.2%) of 56 trials were assessed as low risk of bias(RoB), 38 (67.9%) trials had moderate RoB, and 10 (17.9%) had high RoB; In accordance with the GRADE, the certainty of most evidence was low or very low. Limitations: The overall quality of evidence is low, and the long-term benefits of AAPs are unclear. Conclusion: Adjuvant AAPs had significant efficacy compared with placebo, but treatment decisions need to be made to balance the risks and benefits.