Poor glucose control increases the risk of vascular complications and cardiovascular mortality in patients with diabetes mellitus (DM). Our aim was to evaluate the efficacy of a long-term exercise training program on metabolic control and arterial stiffness in patients with type 2 DM. Fifty men with DM (age 52.3 +/- 5.6 years) were randomly assigned to the exercise training (E) or standard treatment for DM (control [C]) group for 24 months. Supervised exercise training included both endurance and muscle strength training 4 times/week. All exercise sessions were controlled by heart rate and intensity. Glycated hemoglobin A1c, insulin, leptin, blood lipids, blood pressure, maximal oxygen consumption in spiroergometry, and muscle strength were measured every 6 months. Arterial stiffness was assessed by measuring pulse wave velocity. Maximal oxygen consumption in spiroergometry (E 31.9 to 34.8 vs C 32.6 to 31.8 ml/kg/min; p = 0.003), muscle strength (sit-up test, E 12.7 to 20.8 vs C 14.6 to 13.1 times; p <0.001), hemoglobin A1c (E 8.2% to 7.6% vs C 8.0% to 8.3%; p = 0.006), and leptin (E 7.4 to 6.7 vs C 7.4 to 7.9 microg/L; p = 0.013) improved significantly in the E group, but no change or worsening in these variables occurred in the C group. Body weight was not different between groups at 2 years. However, pulse wave velocity increased in both groups (E +0.600 vs C +1.300 m/s; p = 0.27). In conclusion, long-term endurance and strength training was effective and resulted in improved metabolic control of DM compared with standard treatment. Despite significant cardiovascular risk reduction, conduit arterial elasticity did not improve.