Background Acupuncture has been used as a potential therapy for alcohol withdrawal syndrome (AWS), but evidence for its effects on this condition is limited. Objective To assess the effects and safety of acupuncture for AWS. Data sources Central Register of Controlled Trials (CENTRAL), PubMed, Embase, the Cochrane Library, PsycINFO, Chinese Biomedicine Literature (CBM), China National Knowledge Infrastructure (CNKI) and Wan-Fang Database were searched from their inception to August 2016. Study eligibility criteria Randomised controlled trials (RCTs) of drug plus acupuncture or acupuncture alone for the treatment of AWS were included. Data collection and analysis Continuous data were expressed as mean difference (MD) with 95% confidence intervals (95% CI). Dichotomous data were expressed as risk ratio (RR) with 95% CI. Results Eleven RCTs with 875 participants were included. In the acute phase, two trials reported no difference between drug plus acupuncture and drug plus sham acupuncture in the reduction of craving for alcohol; however, two positive trials reported that drug plus acupuncture was superior to drug alone in the alleviation of psychological symptoms. In the protracted phase, one trial reported acupuncture was superior to sham acupuncture in reducing the craving for alcohol, one trial reported no difference between acupuncture and drug (disulfiram), and one trial reported acupuncture was superior to sham acupuncture for the alleviation of psychological symptoms. Adverse effects were tolerable and not severe. Conclusion There was nosignificant difference between acupuncture (plus drug) and sham acupuncture (plus drug) with respect to the primary outcome measure of craving for alcohol among participants with AWS, and no difference in completion rates (pooled results). There was limited evidence from individual trials that acupuncture may reduce alcohol craving in the protracted phase and help alleviate psychological symptoms; however, given concerns about the quantity and quality of included studies, further large-scale and well-conducted RCTs are needed. Protocol registration PROSPERO CRD42016039862.