The present study aimed to investigate whether substrate metabolism, appetite and feeding behaviour differed between high and low energy turnover conditions. Thirteen overweight premenopausal women completed two 1 d trials: low energy turnover (LET) and high energy turnover (HET), in a randomised, cross-over design. In LET, subjects consumed a test breakfast (49 % carbohydrate, 37 % fat, 14 % protein) calculated to maintain energy balance over a 6 h observation period, during which metabolic rate and substrate utilisation were measured and blood samples taken. Immediately following this an ad libitum buffet meal was provided. HET was identical to LET, except that subjects walked on a treadmill for 60 min at 50 % VO2max before the test breakfast, which was increased in size (by about 65 %) to replace the energy expended during the walk and maintain energy balance over the observation period. Postprandial fat balance (i.e. the difference between fat intake and oxidation) was lower and carbohydrate balance higher in HET compared with LET throughout the postprandial period (P < 0·05 for both). After the buffet meal, carbohydrate balance did not differ between trials but energy and fat balances were lower (by 0·28 MJ and 11·6 g, respectively) in HET compared with LET (P < 0·001 for both). Carbohydrate balance immediately before the buffet meal correlated negatively with buffet energy intake (r − 0·49) and postprandial acylated ghrelin responses (r − 0·48), and positively with postprandial glucose responses (r 0·49). These findings demonstrate that HET resulted in a more positive carbohydrate balance than LET, which associated with lower subsequent energy intake. This may have implications for the regulation of body weight.