Objective: To provide the latest evidence on the efficacy and safety of Arbidol (Umifenovir) in COVID-19. Methods: A comprehensive systematic search of the evidence was carried out in PubMed, Cochran library, Embase, and Medrxiv up to October 1, 2020. The Cochrane risk of bias tool and Newcastle-Ottawa Scale checklist were used for assessing the quality of studies. Meta-analysis was performed using RevMan (version 5.3). Results: Fifteen studies were met for the inclusion. No significant difference was observed between Arbidol and control groups in terms of primary outcomes, including negative rate of PCR (NR-PCR) on 7 days (risk ratio [RR] 0.89; P=0.37) and 14 days (RR: 1.10;P=0.17), negative conversion time (NCT) (mean difference [MD]: 0.74; P=0.37), and as well as secondary outcomes (P<0.05). Compared with LPV/r, Arbidol showed a better efficacy in terms of NR-PCR on 14 days (P=0.02). In contrast, NCT in LPV/r was higher (P=0.007). However, not significant difference was found in terms of NR-PCR on 7 days (P=0.05). Adding Arbidol to LPV/r was led to a better efficacy in terms of NR-PCR on 7 days and NCT (P<0.05). Nevertheless, it was not significant reading NR-PCR on 14 days (P=0.99). There is no significant difference Arbidol vs. Interferon /Arbidol and IFN/Arbidol vs. Interferon (P<0.05). Conclusion: Arbidol was not superior to control against COVID-19. Additionally, not major treatment effect was found compared with other therapeutic agents. There are needed well-designed studies with large sample size to establish on efficacy and safety of Arbidol.