This study was designed to assess the effect of physical training on the ventricular beta-adrenergic receptor adenylate cyclase system of diabetic rats. Mild diabetes mellitus was induced by an intravenous (IV) injection of streptozotocin (45 mg/kg). Rats were randomized into a group submitted to a progressive 10-week running program on a treadmill, while another group was kept sedentary. A group of sedentary nondiabetic rats was used as normal controls. Results showed a similar reduction in the density of beta-adrenergic receptors in sedentary diabetic (P less than .05) and trained diabetic rats (P less than .01) compared with controls, without any significant alteration in the dissociation constant. The basal and the sodium fluoride-stimulated maximal adenylate cyclase activities were similar in the three groups. However, the maximal response of adenylate cyclase to isoproterenol was significantly reduced in the two diabetic groups compared with controls (P less than .01). The decrease in adenylate cyclase response to isoproterenol observed in the diabetic groups appeared to be associated with a reduction in the total number of beta-adrenergic receptors and more specifically in those existing in the high-affinity state. On the other hand, the hyperglycemia and hyperglucagonemia present in sedentary diabetic rats was improved by training. These data suggest that the beneficial effects observed in response to training in experimental diabetes are not associated with changes in beta-adrenergic receptor adenylate cyclase system on membranes from ventricular tissue.