The Predicted Mean Vote (PMV) model for thermal comfort determination is unsuitable for use as a generalized index in tropical climates where the weather is hot and humid throughout the year. The current study is aimed to address the discrepancy in the thermal comfort models to estimate the thermal sensation in a tropical climate. Then, a case study was performed on a single-story office building in Malaysia, which has a typical tropical climate, to develop a new modified adaptive Predictive Mean Vote (maPMV) that suits the tropical climate. Experimental measurements were conducted with two groups of occupants. The adoptive mean vote showed that the participants were likely to feel comfortable when the indoor temperature was between 24°C and 26°C. The PMV model overestimated the thermal sensation in the room. Accordingly, a maPMV model was developed to mitigate the inaccuracy of the existing PMV model. The coefficients of the developed model were determined based on the adaptive approach using actual thermal sensation data obtained from a questionnaire survey of the occupants. The overall thermal comfort estimation could be improved using the developed model to reduce the cooling load and the building's energy consumption in tropical areas. © 2023 Wit Press. All rights reserved.