Diabetes is the most important cause of peripheral neuropathy (DPN). No definitive treatment for DPN has been established, and very few data on the role of exercise training on DPN have been reported. We sought to examine the effects of long-term exercise training on the development of DPN in both Types 1 and 2 diabetic patients. Seventy-eight diabetic patients without signs and symptoms of peripheral DPN were enrolled, randomized, and subdivided in two groups: 31 diabetic participants [15 f, 16 m; 49+/-15.5 years old; body mass index (BMI)=27.9+/-4.7], who performed a prescribed and supervised 4 h/week brisk walking on a treadmill at 50% to 85% of the heart rate reserve (exercise group: EXE), and a control group of 47 diabetic participants (CON; 24 f, 23 m; 52.9+/-13.4 years old; BMI=30.9+/-8.4). Vibration perception threshold (VPT), nerve distal latency (DL), nerve conduction velocity (NCV), and nerve action potential amplitude (NAPA) in the lower limbs were measured. We found significant differences on Delta (delta) in NCV for both peroneal and sural motor nerve between the EXE and CON groups during the study period (P<.001, for both). The percentage of diabetic patients that developed motor neuropathy and sensory neuropathy during the 4 years of the study was significantly higher in the CON than the EXE group (17% vs. 0.0%, P<.05, and 29.8% vs. 6.45%, P<.05, respectively). In addition, the percentage of diabetic patients who developed increased VPT (25 V) during the study was significantly higher in the CON than the EXE group (21.3% vs. 12.9%, P<.05). Change on Hallux VPT from baseline to the end of the study was significantly different between the EXE and CON groups (P<.05); no significant change in Malleolus VPT between the two groups occurred. This study suggests, for the first time, that long-term aerobic exercise training can prevent the onset or modify the natural history of DPN.