BACKGROUND: Platelet-rich plasma (PRP) is increasingly used for knee osteoarthritis, but the optimal dosage still needs to be determined. This systematic review and network meta-analysis aimed to compare the efficacy of various PRP doses in treating knee osteoarthritis. METHODS: We searched published data in Embase, PubMed, Scopus, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov, with no language restrictions from the inception to 30 September 2024. We enrolled randomized controlled trials (RCTs) that compared the clinical efficiency of different doses of PRP injection in patients with knee osteoarthritis. The outcomes were reduction in the Visual Analogue Scale (VAS) pain score or improvement of the total Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score. We finally extracted data for four groups, including single PRP injection (PRP1), two PRP injections (PRP2), and three PRP injections (PRP3). We carried out network meta-analyses with the frequentist approach, using a random-effects model to pool the data. RESULTS: A total of 10 eligible RCTs were included, comprising 719 patients. In the 1st month of follow-up, PRP2 and PRP3 demonstrated significantly better VAS and WOMAC scores compared to PRP1, with PRP3 being the most effective in both measures. By months 3 and 6, PRP3 continued to show superior efficacy in both outcomes. PRP2 also exhibited significant improvement in the WOMAC score compared to PRP1 at months 1 and 3. No significant differences were found in the VAS pain score between PRP1 and PRP2 at months 3 and 6. Finally, no major adverse events leading to treatment discontinuation were reported for any PRP groups. CONCLUSION: This network meta-analysis highlights the superior efficacy of higher-dose PRP, particularly PRP3, in reducing pain and improving function in patients with knee osteoarthritis.