Objective. This systematic review and meta-analysis were performed to investigate the efficacy and safety of Chinese herbal medicine (CHM) in the treatment of knee osteoarthritis (KOA). Methods. An electronic search was conducted in eight databases (PubMed, EMBASE, Web of Science, Cochrane Library, Chinese National Knowledge Infrastructure, Chinese Biomedical Literature Database, Chinese VIP Database, and Wanfang Database) from inception until December 2019. The risk of bias assessment of the included RCTs was evaluated by Cochrane collaboration’s tool. The inclusion criteria were RCTs that investigated the efficacy and safety of CHM in the treatment of KOA, with no restrictions on publication status or language. The exclusion criteria included nonrandomized or quasi-RCTs, no clear KOA diagnostic approach, combined Chinese medicinal herbs with other traditional Chinese medicine treatment modalities, and published using repeated data and missing data. We computed the relative risk (RR) and the standard mean difference (SMD) for dichotomous outcomes and continuous outcomes, respectively. When heterogeneity was detected or there was significant statistical heterogeneity (
P
<
0.05
or
I
2
>
50
%
), a random-effects model was employed, followed by further subgroup analysis and metaregression estimations to ascertain the origins of heterogeneity. Otherwise, we used a fixed-effects model (
P
≥
0.05
or
I
2
≤
50
%
). The primary outcome measures were visual analog score (VAS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Lysholm score, and Lequesne index. Secondary outcome measures were the total clinical effective rate and adverse events. The meta-analysis was performed using the Stata 14.0 software. Results. A total of 56 RCTs comprising 5350 patients met the inclusion criteria. This meta-analysis showed that application of CHM as adjuvant therapy or monotherapy for KOA can significantly decrease VAS, WOMAC, and the Lequesne index and improve the Lysholm score as well as the total effective rate. In addition, this treatment has fewer adverse effects, suggesting that CHM is generally safe and well tolerated among patients with KOA. Conclusion. Our study offers supportive evidence that CHM, either adjuvant therapy or monotherapy, reduces the VAS, WOMAC, and Lequesne index and improves the Lysholm score and overall effective rate in patients with KOA. Additionally, CHM was well tolerated and safe in KOA patients. We found frequently used CHMs that might contribute to the formulation of a herbal formula that could be considered for further clinical use. However, given the heterogeneity and limited sample size in this study, larger multicenter and high-quality RCTs are needed to validate the benefits of CHM in the treatment of KOA.