The purpose of this paper is to improve the accuracy of the COMFA outdoor thermal comfort model for application to subjects performing physical activity. A sensitivity analysis was performed to identify conditions where the COMFA model produced erroneous estimates of the heat and moisture exchanges between the human body and the ambient environment, based on data from subjects performing moderate-to-vigorous physical activity. Errors occurred at high metabolic rates (> 400 W m(-2)), high wind speeds (> 4 m s(-1)) and warm air temperatures (> 28 degrees C). Revisions to the clothing resistance (r(c)), clothing vapour resistance (r(c upsilon)), skin tissue resistance (r(t)), and skin temperature (T(sk)) equations were proposed. The revised assessment revealed that subjects had a wide range of thermal acceptability (B = -20 W m(-2) to +150 W m(-2)), which was offset to the warm-end of the comfort scale. The revised model (COMFA*) performed well, predicting the actual thermal sensation of subjects in approximately 70% of cases. This study effectively integrated current empirical research related the effect of wind and activity on the clothing microclimate to improve the application of an outdoor thermal comfort model for subjects performing physical activity.