Objective To evaluate the efficacy and safety of testosterone supplementary treatment for the middle-aged and the senile with insulin resistance (IR). Methods Such databases as PubMed (Jan. 1966 to July 2010), EMbase (Jan. 1984 to July 2010), The Cochrane Library (Issue 3, 2010), CBM (1978 to July 2010), CNKI (Jan. 1994 to July 2010), WanFang Data (1994 to July 2010) and VIP Data (1989 to July 2010) were searched. Randomized controlled trials (RCTs) about testosterone treatment for IR were included. Two reviewers independently extracted the data and evaluated the quality of the included studies. Meta-analyses were performed for the results of homogeneous studies by using RevMan 5.0 software, and other results not suitable for meta-analysis were described with qualitative analyses. Results Nine RCTs involving 573 patients were included. Of them, 308 cases were in the testosterone group and 265 in the placebo group. The baseline data of studies was comparable. The results of meta-analyses showed that, a) Efficacy: testosterone was superior to placebo in decreasing insulin resistance index (HOMA-IR) (WMD= –?0.56, 95%CI –?0.75 to –?0.37) and fasting insulin (FINS) (WMD= –2.4, 95%CI –3.25 to –1.56); and b) Safety: no significant difference was found in prostate specific antigen (PSA) (WMD= –?0.02, 95%CI –?0.22 to 0.18). Conclusion The testosterone supplementary treatment for insulin resistance is superior to the placebo, and there is no significant difference in PSA compared to the placebo. More multicenter double-blind RCTs in large-scale are required to verify this conclusion because of lack of high quality literature with large sample size.