The aim of this study was to compare the effects of two different carbohydrate sources (fructose and glucose) on the metabolic responses in resting and exercising horses. The following regimes were fed in randomized order to five trained horses at rest and immediately before or during exercise. The resting regime comprised 0.6 kg grass meal pellets (control) or 0.6 kg grass meal pellets supplemented with either 50% glucose or 50% fructose. The exercise regime comprised 0.3 kg grass meal (control) or 0.6 kg grass meal pellets supplemented with either 50% glucose or 50% fructose fed immediately before or during simulated endurance exercise on a treadmill (30 km, total running time 120 min; 15 min rest after 60 min running time). Blood samples were collected for the analysis of glucose, insulin, free fatty acids (FFA) and lactate. In resting horses, glucose supplementation resulted in a significantly higher glycaemic and insulinaemic response than the control or fructose feedings (treatment P < 0.05). Plasma glucose levels were significantly higher after glucose supplementation given before or during exercise. Similar plasma glucose concentrations were observed after fructose was fed before exercise, whereas fructose supplementation during exercise resulted in a lower glycaemic response than glucose feeding (P < 0.05). Plasma insulin, FFA and lactate levels showed exercise-related changes (time P < 0.05), but treatment did not effect these results. Plasma glucose concentrations were elevated during the simulated endurance exercise after glucose feeding, and no counter-regulation by insulin occurred. The higher glycaemic response might be beneficial as exogenous glucose can be used as an energetic substrate during prolonged exercise. Fructose exhibited no clear advantages in comparison with glucose as a carbohydrate source for exercising horses.