Background: Previous clinical trials indicated that duloxetine may be effective in the treatment of osteoarthritis (OA) pain. This meta-analysis is conducted to evaluate short term analgesic effect and safety of duloxetine in the treatment of OA. Methods: Electronic databases were searched in February 2019, including PUBMED, EMBASE, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, Web of Science. All eligible studies should be randomized controlled trials (RCTs) comparing duloxetine treatment group to placebo about OA pain relief and safety outcomes. Results: Five RCTs with 2059 patients were involved in this systematic review and meta-analysis. Compared to placebo, duloxetine treatment showed significant better result, with higher reduction pain intensity (mean difference [MD] = –0.77, P < .00001), higher rates of both 30% and 50% reduction in pain severity (risk ratio [RR] = 1.42, P < .00001; RR = 1.62, P < .00001), lower mean Patient Global Improvement-Inventory (PGI-I) score (MD = –0.48, P < .00001). The results of the Western Ontario and McMaster Universities (WOMAC) score change from baseline to endpoint also favored duloxetine treatment group in all four categories, including total (MD = –5.43, P < .00001), pain (MD = –1.63, P = .001), physical function (MD = –4.22, P < .00001), and stiffness score (MD = –0.58, P < .00001). There were higher rates of treatment-emergent adverse events (TEAEs) (RR = 1.32, P < .00001) and discontinuation (RR = 1.88, P < .00001) in duloxetine group. However, there was no significant difference in the incidence of severe adverse events (SAEs) between these 2 groups (RR = 0.84, P = .68). Conclusion: Duloxetine was an effective and safe choice to improve pain and functional outcome in OA patients. However, further studies are still needed to find out the optimal dosage for OA and examine its long-term efficacy and safety. Trial registration number: CRD42019128862