Objectives This study assessed the effectiveness, compliance, and safety of dextrose prolotherapy for patients with knee osteoarthritis. Data sources PubMed, EMBASE, the Cochrane Library Database, and the Scopus database from their inception to December 31, 2021. Methods This study was conducted in accordance with the guidelines recommended by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis. Randomized controlled trials regarding the effectiveness of dextrose prolotherapy in knee osteoarthritis were identified. The included trials were subjected to meta-analysis. Risk of bias was assessed using the Cochrane risk of bias tool. Subgroup and random-effects metaregression analyses were performed to explore any heterogeneity ( I2) of treatment effects across studies. Results A total of 14 trials enrolling 978 patients were included in the meta-analysis. Compared with placebo injection and noninvasive control therapy, dextrose prolotherapy had favorable effects on pain, global function, and quality of life during the overall follow-up. Dextrose prolotherapy yielded greater reductions in pain score over each follow-up duration than did the placebo. Compared with other invasive therapies, dextrose prolotherapy generally achieved comparable effects on pain and functional outcomes for each follow-up duration. Subgroup results indicated that combined intra-articular and extra-articular injection techniques may have stronger effects on pain than a single intra-articular technique. Conclusions Dextrose prolotherapy may have dose-dependent and time-dependent effects on pain reduction and function recovery, respectively, in patients with knee osteoarthritis. Due to remarkable heterogeneity and the risk of biases across the included trials, the study results should be cautiously interpreted.