BACKGROUND: Findings on the usefulness of massage therapy (MT) in postoperative pain management are often inconsistent among studies. OBJECTIVES: This study's aim is to conduct a meta-analysis of randomized controlled trials (RCT) to clarify the effects of massage therapy in the treatment of postoperative pain. METHODS: Three databases (PubMed, Embase, and Cochrane Central Register of Controlled Trials) were searched for RCTs published from database inception through January 26, 2021. The primary outcome was pain relief. The quality of RCTs was appraised with the Cochrane Collaboration risk of bias tool. The random-effect model was used to calculate the effect sizes and standardized mean difference (SMD) with 95 % confidential intervals (CIs) as a summary effect. The heterogeneity test was conducted through I RESULTS: The analysis included 33 RCTs and showed that MT is effective in reducing postoperative pain (SMD, -1.32; 95 % CI, -2.01 to -0.63; p = 0.0002; I LIMITATIONS: Publication bias is possible due to the inclusion of studies in English only. Additionally, the included studies were extremely heterogeneous. Double-blind research on MT is difficult to implement, and none of the included studies is double-blind. There was some heterogeneity and publication bias in the included studies. In addition, there is no uniform evaluation standard for the operation level of massage practitioners, which may lead to research implementation bias. CONCLUSIONS: MT is effective in reducing postoperative pain in both short and long terms.