Purpose: Rheumatoid arthritis (RA) is a pervasive inflammatory autoimmune disease that seriously impairs human health and requires more effective non-pharmacologic treatment approaches. This study aims to systematically review and evaluate the efficacy of yoga for patients with RA. Methods: Medline (through PubMed), Cochrane Library, EMBASE (through SCOPUS), and Web of Science database were screened through for articles published until 20 July 2020. Randomized controlled trials (RCTs) of yoga in patients with RA were included. Outcomes measures were pain, physical function, disease activity, inflammatory cytokines, and grip strength. For each outcome, standardized mean differences (SMD) and 95% confidence intervals (CI) were calculated. Result: Ten trials including 840 patients with RA aged 30–70 years were identified, with 86% female participants. Meta-analysis revealed a statistically significant overall effect in favor of yoga for physical function (HAQ-DI) (5 RCTs; SMD = −0.32, 95% CI −0.58 to −0.05, I2 = 15%, P = 0.02), disease activity (DAS-28) (4 RCTs; SMD = −0.38, 95% CI −0.71 to −0.06, I2 = 41%, P = 0.02) and grip strength (2 RCTs; SMD = 1.30, 95% CI 0.47–2.13, I2 = 63%, P = 0.002). No effects were found for pain, tender joints, swollen joints count or inflammatory cytokines (i.e., CRP, ESR, IL-6, and TNF-α). Summary: The findings of this meta-analysis indicate that yoga may be beneficial for improving physical function, disease activity, and grip strength in patients with RA. However, the balance of evidence showed that yoga had no significant effect in improving pain, tender joints, swollen joints count, and inflammatory cytokines in patients suffering from RA. Considering methodological limitations, small sample size, and low-quality, we draw a very cautious conclusion in the results of the estimate of the effect. High-quality and large-scale RCTs are urgently needed in the future, and the real result may be substantially different. © Copyright © 2020 Ye, Chen, Shen, Chen and Xu.