Objective To systematically evaluate the effectiveness and safety of berberine in the treatment of type 2 diabetes. Methods The databases including The Cochrane Library, PubMed (1966 to October 2011), Excerpta Medica Database (EMbase, 1974 to October 2011), Chinese National Knowledge Infrastructure databases (CNKI, 1994 to October 2011), the Chinese Scientific and Technical Journals database (VIP, 1989 to October 2011), and China Doctor Dissertation Full-text Database (CDFD, 1979 to 2011) and China Master Dissertation Full-text Database (CMFD, 1979 to 2011) were searched. The randomized controlled trials (RCTs) on berberine in the treatment of type 2 diabetes were screened according to the inclusive and exclusive criteria. The data were extracted, the quality was assessed, and the systematic review was conducted by using Revman 5.0 software. Results Ten RCTs involving 647 Chinese patients with DM 2 were included, and the quality of each study was generally low. The interventions in the treatment groups were berberine or combined with metformin or glipizide. The control groups included placebo, lifestyle intervention, pioglitazone, rosiglitazone or metformin. Because the experiment and control groups in each included trials were different in drug type and dose, disease duration, and treatment regimens, only the results of all trials were reported rather than performing Meta-analysis. The berberine group was superior to the placebo and lifestyle intervention groups in lowering fasting blood glucose (FBG), but it was not so obviously effective in lowering the postprandial blood glucose (PBG), hemoglobin A1c and BMI and regulating lipid metabolism compared with the placebo, lifestyle intervention, and western hypoglycemic agents. In addition, the berberine treatment had no side effects of hypoglycemia although a few patients complained of gastrointestinal adverse reaction, and there was no significant difference when compared with the placebo and lifestyle intervention groups. Conclusion Berberine is effective in lowering FBG, but not better than metformin, glipizide and rosiglitazone. It is undefined in decreasing PBG, HbA1c, BMI and regulating lipid metabolism, and it will not lead to hypoglycemia except for a few and mild gastrointestinal adverse effects. The current clinical studies on berberine for DM 2 are low in methodology and reporting quality, which has to be further proved by more high-quality clinical trails.