Indoor thermal climate is an important issue affecting the health and productivity of building occupants. In the designing of commercial air-conditioning systems, it is believed that the conventional fixed temperature set point concept is limited because indoor comfort temperature depends on the business culture, such as the nature of activities and dress code of occupants, etc. Researchers have been interested in investigating adaptive temperature control for a realistic in-situ control of comfort. Unfortunately, those studies put great emphasis on energy saving opportunities and sometimes might result in thermal discomfort to individuals. This study argues that complaints of thermal discomfort from individuals, despite representing only a small portion of the population, should not be ignored and can be used to determine the temperature setting for a population in air-conditioned environment. In particular, findings of a new notion of Bayesian adaptive comfort temperature (BACT) in air-conditioned buildings in a humid and subtropical climate like Hong Kong are reported, and the adaptive interface relationship between occupants’ complaints of thermal discomfort and indoor air temperature is determined. This BACT algorithm is intended to optimise the acceptance of thermal comfort, as determined by physical measurements and subjective surveys.