Background: Plant-derived therapies are traditionally used as medicines, but they have generally not been studied with the same rigor as pharmaceutical agents. This review summarizes the use of plant-derived products for osteoarthritis. Methods: Sixty-three identified trials were summarized for pain, function, and safety outcomes using standardized mean differences (SMDs) and relative risks. Results: Plant-derived therapies are effective for treating pain compared to placebo, as assessed using visual analog scores and numerical rating scales (SMD, 1.08; 95% confidence interval [CI]: 0.72-1.44), or Western Ontario and McMaster University Osteoarthritis Index (WOMAC)/Knee injury and Osteoarthritis Outcome Score (KOOS) pain scales (SMD, 0.98; 95% CI: 0.62-1.35). Classes demonstrating overall efficacy in more than one trial for either visual analog scores or WOMAC pain included Boswellia serrata, capsaicin, and ginger; there was single-trial evidence of the efficacy of another nine agents. Plant-derived therapies have similar efficacy to an active comparator (SMD, 0.32; P=0.08; -0.08; P=0.14). Therapies are also effective for functional outcomes compared to placebo (SMD, 0.92; P,0.001). However, significant heterogeneity remains for all pain and function outcomes, indicating that the results need to be interpreted with caution. Risk of adverse events was similar to placebo (relative risk =1.13; P=0.1), but reduced compared to an active comparator (relative risk, 0.75; P,0.001). Conclusion: Plant-derived therapies may be efficacious in treating osteoarthritic pain and functional limitations, and they appear to be safer than other active therapies. However, quality trials and long-term data are lacking, and the number of trials for each therapy is limited. Comparisons would be assisted by trial standardization.