Conventional models of building occupants' environmental preferences such as thermal comfort are used to give guidelines for the average environmental conditions that will satisfy large groups of people. The research described in this paper investigates how the preferences of an individual occupant can be modeled to predict their preferred thermal and environmental conditions. A novel, Internet based questionnaire was developed to gather thermal sensation votes. Using 361 subjective responses various models were used to predict the thermal sensation vote of the occupant and the results were compared with the Predicted Mean Vote (PMV) and the actual vote. Preliminary results show that Neuro-Fuzzy model produces better results than PMV, Parametric and Fuzzy models.