In people with type 1 diabetes mellitis (DM), insulin administration, food intake, and exercise have to be carefully matched to avoid either hypo- or hyperglycemia. People with type 2 DM have some insulin secretion, which changes with needs. Accordingly, during exercise, these people do not run the same metabolic risks as people with type 1 DM. However, a contraction-mediated increase in glucose clearance in muscle is intact in type 2 DM. Therefore, in the postabsorptive state in diet-treated type 2 DM, a marked reduction in hyperglycemia can occur during prolonged moderate exercise. Sulfonylurea drugs augment the rate of decline in plasma glucose, because stimulation of insulin secretion reduces hepatic glucose production. After abstention from sulfonylurea for 5 days, the rate of decrease in plasma glucose with exercise is also enhanced, but from a higher glucose level. In the postabsorptive state, brief vigourous exercise elicits an increase in plasma glucose concentration, reflecting an exaggerated counterregulatory hormone response and glucose production. Moreover, insulin sensitivity is reduced in the early postexercise period. In the postprandial state, both prolonged moderate exercise and intermittent high-intensity exercise markedly decrease meal-induced increases in glucose, insulin, and C-peptide concentrations, whereas glucose appearance in plasma is unchanged. When exercise bouts are isocaloric, responses are identical, indicating that overall energy expenditure, and not peak exercise intensity, is the major determinant of exercise-induced changes in overall glucose homeostasis and insulin secretion in type 2 DM. Neither prolonged moderate nor intermittent high-intensity exercise performed in the postprandial state influences glucose or insulin responses to a subsequent meal. Finally, in people with type 2 DM, after a high-fat meal, prolonged moderate exercise reduces the exaggerated increases in plasma concentrations of triglycerides contained in chylomicrons and very low-density lipoproteins.