Short-term elevations of stress hormones cause an increase in glycemia. However, the effect of intermittent stress on development of type 2 diabetes mellitus is unclear. We hypothesized that recurrent intermittent restraint stress would deteriorate glycemia. Male, prediabetic Zucker diabetic fatty (ZDF) rats were restrained 1 hour per day, 5 days per week for 13 weeks and compared with unstressed, age-matched diabetic controls and lean nondiabetic rats. To differentiate the effects of recurrent restraint stress per se vs restraint-induced inhibition of food intake, a pair-fed group of rats was included. Surprisingly, recurrent restraint and pair feeding delayed fed and fasting hyperglycemia, such that they were lowered 50% by restraint and 30% by pair feeding after 13 weeks. Rats that were previously restrained or pair fed had lower glucose levels during a glucose tolerance test, but restraint further improved the return of glucose to baseline compared to pair feeding (P<.05). This was despite pair-fed rats having slightly lowered food intake and body weights compared with restrained rats. Restraint and pair feeding did not alter insulin responses to an intraperitoneal glucose tolerance test (IPGTT) or fasting insulin, and did not lower plasma lipids. Interestingly, restraint normalized basal corticosterone to one third that in control and pair-fed rats, prevented increases in pretreatment corticosterone seen with pair feeding, and led to habituation of restraint-induced corticosterone responses. After 13 weeks of treatment, multiple regression analysis showed that elevations in basal corticosterone could explain approximately 20% of the variance in fed glucose levels. In summary, intermittent restraint and its adaptations delayed hyperglycemia and improved glucose control in Zucker diabetic fatty rats. These benefits can be partially explained by restraint-induced lowering of food intake, but additional improvements compared to pair feeding may involve lower overall corticosterone exposure with repeated restraint. Paradoxically, these novel investigations suggest some types of occasional stress may limit development of diabetes.