The aim of this study was to assess the effects and safety of salicylates on type 2 diabetes mellitus (T2DM). We searched six databases (Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, CBM, CNKI and VIP) for all randomized controlled trials (RCTs) and self-control studies which investigated the effects of salicylates on T2DM. We included 34 RCTs and 17 self-control studies involving 13 464 patients with T2DM. It was demonstrated that salicylates had obvious effects on several parameters for patients with T2DM. (1) Any dose of salicylates could significantly reduce HbA1c level [mean difference (MD) -0.39%; 95% CI -0.47 to -0.32] in RCTs, but only high doses of salicylates (a parts per thousand yen3000 mg/day) could effectively reduce fasting plasma glucose (FPG) level [standardized mean difference (SMD) -1.05; 95% CI -1.47 to -0.62] for patients with T2DM in both RCTs and self-control studies. Furthermore, high doses of salicylates could also increase plasma fasting insulin level (MD 12.20 mU/L; 95% CI 3.33 to 21.07); (2) In both RCTs and self-control studies, high doses of salicylates could significantly reduce plasma triglycerides concentration. The results for RCTs were MD -0.44 mmol/L, 95% CI -0.71 to -0.18, and those for self-control studies were 227 +/- 29 mg/dL (pre-treatment) and 117 +/- 8 mg/dL (post-treatment) (P=0.009); (3) All trials which reported cardiovascular events were RCTs using low doses (< 1000 mg/day) of salicylates, and it was revealed that aspirin could significantly reduce the risk of myocardial infarction (OR 0.73; 95% CI 0.57 to 0.92); (4) Two RCTs and two self-control studies with a parts per thousand yen3000 mg/day salicylates reported adverse effects, and the overall effects were mild, and tinnitus occurred most frequently. No evidence of gastrointestinal bleeding was found in all these studies. In conclusion, from our systematic review, the anti-diabetic effect of salicylates is in a dose-dependent manner. High doses of salicylates may have beneficial effects on reducing FPG, HbA1c level and increasing fasting insulin concentration, and may also have some positive effects on lipidemia and inflammation-associated parameters for patients with T2DM, without serious adverse effects.