Objective To assess the effectiveness and safety of breviscapine on diabetic nephropathy. Methods All randomized or quasi-randomized controlled trials of breviscapine on diabetic nephropathy were performed. All of the clinical trials were searched from the Cochrane Controlled Trials Registered, Medline, Embase, National Knowledge Infrastructure Database, the Chinese Biological Medicine Database, the Chinese Science and Technology Journal Full-text and the references of all included trials. The selection of studies, assessment of methodological quality and data extraction were performed independently by two reviewers according to predefined inclusion and exclusion criteria. Results Thirty-three clinical trials including 2 322 patients of diabetic nephropathy met the inclusion criteria. But most included trials were of low quality and small sample. Until now, there were no clinical trials with multicentre, large sample and high quality. A “Funnel plot” showed asymmetry, which indicated possible publication bias and low quality in methodology. And publication bias showed that the trials with negative results might not be published. The results of meta-analysis indicated that: 1. Breviscapine showed more effects on the decrease of the 24-hour urinary albumin excretion rate (UAER), 24-hour urinary protein, serum creatinine (Scr), total cholesterol, triglyceride, plasma viscosity and fibrinogen. 2. Breviscapine showed less effect on the decrease of the 24-hour urinary protein when compared to angiotensin-converting enzyme inhibitor, it seemed as same effective as ACEI on decrease of 24-hour urinary albumin excretion rate (UAER), serum creatinine (Scr) and blood urea nitrogen (BUN); 3. Breviscapine showed more effect on the decrease of 24-hour urinary protein and fibrinogen when compared to other Chinese herbal medicine (Salvia miltiorrhiza); 4. Breviscapine showed less effect on decrease of the 24-hour urinary albumin excretion rate (UAER) when compeard to Prostaglandin E1. No significant adverse effects were reported. Conclusion Breviscapine shows some effects and relatively safe on diabetic nephropathy. However, the evidence is not b enough because of some of the low-quality trials and publications bias. Rigorous designs, randomized, double-blind, placebo-controlled trials of Breviscapine for diabetic nephropathy are needed to further assess the effect.