The present study examined the effects of a 3-week fish-oil supplementation (6 g/d) on the rate of plasma glucose disappearance (Rd glucose), hepatic glucose production (HGP), carbohydrate oxidation and lipid oxidation during exercise. Six untrained males (23±1 years; 67·6±2·7kg) performed two 90min cycling exercise sessions at 60% of maximal O2 output separated by 20 d. During the 20 d before the first test, they ingested 6g olive oil/d, then 6g fish oil/d during the 20 d before the second test. Plasma glucose fluxes and lipolysis were traced using 6,6-[2H2]glucose and 1,1,2,3,3-[2H5]glycerol respectively. Substrates oxidation was obtained from indirect calorimetry. At rest HGP and the Rd glucose were similar after olive oil and fish oil (1.83 (se 0·05) v. 1·67 (se 0·11) mg/kg per min). During exercise, fish oil reduced the stimulation of both the Rd glucose (5·06 (se 0·23) v. 6·37 (se 0·12) mg/kg per min; P<0·05) and HGP (4·88 (se 0·24) v. 5·91 (se 0·21) mg/kg per min; P<0·05). Fish oil also reduced glucose metabolic clearance rate (6·93 (se 0·29) v. 8·30 (se 0·57) ml/min). Carbohydrate oxidation tended to be less stimulated by exercise after fish oil than after olive oil (12·09 (se 0·60) v. 13·86 (se 1·11) mg/kg per min; NS). Lipid oxidation tended to be more stimulated by exercise after fish oil (7·34 (se 0·45) v. 6·85 (se 0·17) mg/kg per min; NS). Glycaemia, lactataemia, insulinaemia and glucagonaemia were similarly affected by exercise after fish oil and olive oil. Lipolysis at rest was similar after fish oil and olive oil (2·92 (se 0·42) v. 2·94 (se 0·28) μmol/kg per min) and similarly stimulated by exercise (6·42 (se 0·75) v. 6·77 (se 0·72) μmol/kg per min). It is concluded that fish oil reduced the Rd glucose by 26% by reducing glucose metabolic clearance rate, possibly by facilitating fat oxidation, and reduced HGP by 21%, possibly by a feedback mechanism.