OBJECTIVE： To systematically review the efficacy and safety of Xiaoke pill versus glibenclamide in the treatment of type 2 diabetes mellitus， and provide evidence-based reference for clinical treatment. METHODS：Retrieved from Medline， PubMed， EMBase， Cochrane Library， CBM，CJFD， Wanfang Database and VIP Database， randomized controlled trials （RCT） about the Xiaoke pill （test group） versus glibenclamide （control group） in the treatment of type 2 diabetes mellitus were enrolled. Meta-analysis was performed by using Rev Man 5.2 software after document selection， data extract and quality assessment by Cochrane systematical evaluation. RESULTS： A total of 15 RCT were included， involving 3 319 patients. Results of Meta-analysis showed the hemoglobin A1c （HbA1c） level [MD＝－0.39，95％CI（－0.75，－0.02），P＝0.04]，fasting plasma glucose （FPG） level [MD＝－0.70，95％CI（－1.27，－0.12），P＝0.02]， 2 h PG[MD＝－0.87， 95％CI（－1.55，－0.20），P＝0.01]， incidence of thirst with desire for drinks [RR＝3.35，95％CI（1.92， 5.85），P＜0.001]， incidence of tiredness and debilitation[RR＝5.74，95％CI（3.52，9.36），P＜0.001] and incidence of hypoglycemia [RR＝0.67，95％CI（0.49， 0.91），P＝0.01] in test group were significantly lower than control group， the differences were statistically significant. CONCLUSIONS： Xiaoke pill has better efficacy and safety than glibenclamide in the treatment of type 2 diabetes mellitus， can obviously improve the HbA1c， FPG，2 h PG tevel and TCM symptoms.