Outdoor thermal environment is affected by variables like air temperature, wind velocity, humidity, temperature of the radiant surfaces, and solar radiation, which can be expressed by a single number - the thermal index. Since these variables are subject to annual and diurnal variations, prediction of thermal comfort is of special importance for people to plan their outdoor activities. The purpose of this research was to develop and apply the extreme learning machine for forecasting physiological equivalent temperature values. The results of the extreme learning machine model were compared with genetic programming and artificial neural network. The reliability of the computational models was accessed based on simulation results and using several statistical indicators. According to obtained results, it can be concluded that extreme learning machine can be utilized effectively in short term forecasting of physiological equivalent temperature.