Patients with non-insulin-dependent diabetes are at greatly increased risk for coronary artery disease. Although exercise training has been shown to decrease risk factors, the presence of obesity, older age, and a sedentary lifestyle make a high-intensity exercise program an unrealistic choice of therapy. Therefore, we examined the effect of a low-to-moderate-intensity (mean 69 per cent of maximal heart rate) walking program on lipids, glucose, insulin, glycosylated hemoglobin and cardiovascular fitness. Nine women and seven men, mean age 56, were randomly assigned to a control or an exercise group which exercised three times per week for two months. Supervised exercise sessions consisted of 40-45 minutes of walking and/or slow jogging. Subjects continued on their usual diets. The trained group showed a significant improvement in VO2max from 1.65 to 1.95 L/min. Resting systolic blood pressure decreased from 141 to 130 mm Hg after training, and resting heart rate decreased from 88 to 81. Glycosylated hemoglobin decreased in the exercise group in seven or eight subjects and in only two of eight controls. Triglycerides decreased in the exercise group from 285 to 223 mg/dl. Body weight, total and HDL cholesterol, glucose, and insulin did not change in either group. These data indicate that a low-to-moderate level of aerobic training, independent of dietary changes, is an effective and feasible method of improving cardiovascular risk factors: physical fitness, systolic blood pressure, plasma triglycerides, and glycemic control in non-insulin-dependent diabetic subjects.