In this study we compared substrate use at submaximal intensities of a maximal graded exercise test (GXT) with that derived from equivalent intensities during continuous submaximal steady-state exercise in obese and normal-weight women. Sedentary obese (n = 20, body fat > 30%) and normal-weight (n = 15, body fat < or =30%) women performed three treadmill tests with concurrent metabolic measurements. Maximal oxygen consumption (VO2max) was determined using the Bruce protocol, followed by two, randomly assigned, continuous 15-min, steady-state exercise bouts, on different days; one bout at 50% and one bout at 75% VO2max. Analysis of variance revealed no significant differences between groups for blood lactate or respiratory exchange ratio (R) values at any point during exercise. Therefore, obese and normal-weight group data were combined for subsequent analyses. The R at 50% VO2max from the GXT [0.83 (0.01)] was significantly (P < 0.05) lower than at 8 min [0.90 (0.01)] and 15 min [0.89 (0.01)] of steady-state exercise, whereas at 75% VO2max, the GXT R [0.96 (0.01)] was similar to that seen at 8 min [0.96 (0.01)] and at 15 min of steady-state exercise [0.93 (0.01)]. Blood lactate values at 50% VO2max were similar between the GXT [1.66 (0.10) mM] and steady-state exercise [1.65 (0.09) mM], but at 75% VO2max the GXT blood lactate values [2.58 (0.21) mM] were lower than after 15 min of steady-state exercise [4.65 (0.46) mM]. Total exercise fat oxidation was greater at 50% compared to 75% VO2max. There was no difference in substrate use between sedentary obese and normal-weight women either at rest or during steady-state exercise at the same relative intensity. Total fat oxidation was greater during low- (50% VO2max) compared to high-intensity (75% VO2max) exercise. Data from a GXT cannot be used to predict R or substrate utilization values for the purpose of exercise prescription.