Knee osteoarthritis (KOA) is a common chronic degenerative disease of the elderly. Electro-acupuncture (EA) is considered as a beneficial treatment for KOA, but the conclusion is controversial. This systematic review compiled the evidence from 11 randomized controlled trials to objectively assess the effectiveness and safety of EA for KOA. Eight databases including PubMed, Cochrane Library, Clinic trials, Foreign Medical Literature Retrial Service (FMRS), Science Direct, China National Knowledge Infrastructure (CNKI), Chinese Scientific Journal Database (VIP), and Wanfang Data were extensively searched up to 5 July 2016. The outcomes included the evaluation of effectiveness, pain and physical function. Risk of bias was evaluated according to the Cochrane risk of bias tool. Eleven RCTs with 695 participants were included. Meta-analysis indicated that EA was more effective than pharmacological treatment (RR [Formula: see text] 1.14; 95% CI [Formula: see text] 1.01,1.28; [Formula: see text]) and manual acupuncture (RR [Formula: see text] 1.12; 95% CI [Formula: see text] 1.02,1.22; [Formula: see text]). Also, EA had a more significant effect in reducing the pain intensity (SMD [Formula: see text]; 95% CI [Formula: see text]; [Formula: see text]) and improving the physical function in the perspective of WOMAC (MD [Formula: see text]; 95% CI [Formula: see text], 5.56; [Formula: see text]) and LKSS (pharmacological treatment: MD [Formula: see text]; 95% CI [Formula: see text], 6.64; [Formula: see text]). Furthermore, these studies implied that EA should be performed for at least 4 weeks. Conclusively, the results indicate that EA is a great opportunity to remarkably alleviate the pain and improve the physical function of KOA patients with a low risk of adverse reaction. Therefore, more high quality RCTs with rigorous methods of design, measurement and evaluation are needed to confirm the long-term effects of EA for KOA.