Objective To systematically evaluate the efficacy and safety of traditional Chinese medicine (TCM) for non-alcoholic steatohepatitis (NASH). Methods Such databases as CNKI, VIP, CBM, and PubMed (from the date of their establishments to November 2009) were searched. Randomized controlled trials of TCM treatment of NASH were included. According to the principle of ZANG-fu Differentiation, the method recommended by the Cochrane Collaboration was applied to evaluate the quality of the studies and to conduct meta-analysis. Results Twenty-two studies were included, of which 18 were of low quality and 4 of high quality. The results of meta-analyses showed: there was no difference between liver-oriented TCM treatment and western medicine treatment in terms of the recovery rate, total effective rate, liver function, and blood lipid; the liver and spleen oriented TCM treatment was better than the western medicine treatment in terms of the recovery rate, total effective rate, liver function, and blood lipid, but similar to that of western medicine on TG and HDL-C; the liver, spleen and kidney oriented TCM treatment was better than the western medicine in terms of the recovery rate, total effective rate, liver function, lipids, but showed no difference on HDL-C; the liver and kidney oriented TCM treatment was better than the western medicine treatment in terms of the recovery rate, total effective rate, liver function, LDL-C, and HDL-C, but showed no differences on TC and TG. Conclusion The TCM has significant effects on NASH. The liver, spleen and kidney oriented TCM treatment is significantly better than the western medicine treatment in terms of the recovery rate, total effective rate, and liver function. TCM has no serious adverse reactions.