Summary Objective To investigate the efficacy of different electrical stimulation (ES) therapies in pain relief of patients with knee osteoarthritis (OA). Method Electronic databases including MEDLINE, Embase and Cochrane Library were searched through for randomized controlled trials (RCTs) comparing any ES therapies with control interventions (sham or blank) or with each other. Bayesian network meta-analysis was used to combine both the direct and indirect evidence on treatment effectiveness. Results 27 trials and six kinds of ES therapies, including high-frequency transcutaneous electrical nerve stimulation (h-TENS), low-frequency transcutaneous electrical nerve stimulation (l-TENS), neuromuscular electrical stimulation (NMES), interferential current (IFC), pulsed electrical stimulation (PES), and noninvasive interactive neurostimulation (NIN), were included. IFC is the only significantly effective treatment in terms of both pain intensity and change pain score at last follow-up time point when compared with the control group. Meanwhile, IFC showed the greatest probability of being the best option among the six treatment methods in pain relief. These estimates barely changed in sensitivity analysis. However, the evidence of heterogeneity and the limitation in sample size of some studies could be a potential threat to the validity of results. Conclusion IFC seems to be the most promising pain relief treatment for the management of knee OA. However, evidence was limited due to the heterogeneity and small number of included trials. Although the recommendation level of the other ES therapies is either uncertain (h-TENS) or not appropriate (l-TENS, NMES, PES and NIN) for pain relief, it is likely that none of the interventions is dangerous. Level of evidence LevelⅡ, systematic review and network meta-analysis of RCTs.