To test the hypothesis that high fructose (HF) consumption divergently affects exercise capability, as a function of feeding duration, rats were fed a normal (as control) or a normal caloric diet with HF for 3, 6, 10 and 30 days respectively, and run on a treadmill. Results show that running distance and work were significantly increased, associated with greater exercise oxygen consumption in rats fed HF for 3 (HF-3D) and 6 days, but were decreased in rats fed HF for 30 days (HF-30D) compared to their respective control rats. Shear stress-induced dilation (SSID) in isolated plantaris muscle arterioles was significantly greater in HF-3D than control rats. The difference in SSID between the two groups was abolished by L-NAME, suggesting an NO-mediated response. Expression of phosphorylated/activated eNOS and release of nitrite/NO were significantly increased in vessels of HF-3D than controls. In contrast, arterioles isolated from the hypertensive HF-30D rats displayed significantly attenuated NO-mediated SSID accompanied with greater production of superoxide compared to vessels of controls. Additionally, the NO-dependent modulation of myocardial oxygen consumption (MVO2) was also impaired in HF-30D rats, which was prevented by blocking superoxide production with apocynin, an inhibitor that also normalized the reduced SSID in HF-30D rats. In conclusion, short-term (3-6 days) HF feeding enhances exercise potential by an increase in endothelial sensitivity to shear stress, which stimulates eNOS to release NO, leading to better tissue perfusion and utilization of oxygen. However, long-term (30 days) HF feeding initiates endothelial dysfunction by superoxide-dependent mechanisms to compromise exercise performance.