The relationship between moderate exercise training (five 45 min sessions/week, brisk walking at 62 beta +/- 2% VO2max for 15 weeks, psychological well-being and mood state was investigated in a group of 35 sedentary, mildly obese women. A 2 (exercise (EX) (N = 18), and nonexercise (NEX) (N = 17) groups) x 3 (baseline, 6-week, 15-week testing sessions) factorial design was used with data analyzed using repeated measures ANOVA. Four psychological tests were administered: Daily Hassles Scale (DHS), General Well-being Schedule (GWB), Spielberger State Anxiety Inventory (S-Anxiety), and Profile of Mood States (POMS). The EX and NEX groups had significantly different patterns of change over time for GWB total scores [F(2,66) = 5.72, p = 0.005] and the GWB subscales 'energy level' and 'freedom from health concern or worry'. Scores for the EX group were elevated at both 6 and 15 weeks. General well-being total scores and subscale 'energy level' scores were significantly correlated with improvement in submaximal cardiorespiratory fitness (r = -0.41, p = 0.014; r = -0.40, p = 0.017, respectively). Exercise training also had a significant effect on frequency but not intensity of DHS scores, and S-Anxiety, with a significant decrease seen in the EX group at 6 weeks but not 15 weeks. Profile of Mood States scores were not significantly related to exercise training. These data support the results of other studies that have reported improvement in general psychological well-being with exercise training.