Insulin resistance is a determining factor in the pathophysiology of type 2 diabetes mellitus (T2DM). Exercise is known to improve insulin resistance, but a systematic review of the literature is lacking. This systematic review and meta-analysis focused on identifying evidence for the effectiveness of a structured exercise intervention program for insulin resistance in T2DM. We searched MEDLINE via PubMed, CINHAL, Scopus and Web of Science, and the Cochrane Central Register of Controlled Trials for reports of studies on fasting insulin, homeostatic model assessment for insulin resistance (Homa-IR), fasting blood sugar, glycated hemoglobin and body mass index in patients with T2DM and healthy controls that were published between 1990 and 2017. Data are reported as the standardized mean difference or mean difference with 95% confidence intervals (CIs). Among 2242 records retrieved, only 11 full-text articles were available for meta-analysis. Data for 846 participants were analyzed, 440 in the intervention group, and 406 in the control group. The mean difference for fasting insulin level was−1.64 (95% CI; −3.38 to 0.10), Homa-Ir 0.14 (−1.48 to 1.76), fasting blood sugar−5.12 (−7.78 to−2.45), hemoglobin A1c 0.63 (−0.82 to 2.08) and body mass index−0.36 (−1.51 to 0.79). The evidence highlights the effectiveness of a structured exercise intervention program for insulin resistance in T2DM with a moderate level 2 of evidence.