The Universal Thermal Climate Index (UTCI) assesses the interaction of ambient temperature, wind, humidity and radiant fluxes on human physiology in outdoor environments on an equivalent temperature scale. Based upon the dynamic thermal sensation (DTS) from the UTCI-Fiala model of human thermoregulation, the UTCI allows for thermal comfort prediction.Here we compare those predictions to thermal sensation votes as recorded on the 7-unit ASHRAE scale for two Brazilian cities, Curitiba and Pelotas. Outdoor comfort surveys from 1551 respondents in Curitiba and 1148 in Pelotas, respectively, yielded negligible bias and less than one unit rootmean square error (rmse), which was similar in magnitude for both study areas. Residual analysis revealed that factors such as age, sex, body composition, site morphology (open space, street canyon), climatic state (comfort/discomfort) and clothing choice only explained a small portion of the prediction error variance, which in the total sample was dominated for over 94% by residual inter-individual variability. Adding historical weather information from the previous three days gave superior information compared to longer time lags and helped to reduce the residual variance to 88%. Those findings underpin current limitations in individual thermal comfort prediction, whereas personal and situational factors hardly affected 2 UTCI predictive performance, which showed reasonable accuracy at the population level.