<p>Automatic heating adjustment as an environmental assessment technique has long been a critical challenge not just for improvement of fuel efficiency of heating appliances but also for human comfort for ease of use of those appliances. Traditional heating adjustment has mainly been made manually due to a lack of comprehensive user management and individual thermal comfort. Thermal comfort is to achieve a comfortable indoor temperature that people perceive as neither too hot nor too cold as desired and individual thermal comfort itself can be directly measured as thermal sensation vote (TSV) or indirectly assessed as predicted mean vote (PMV) using a complex equation. As latest heating management combining internet of things (IoT) technology with heating systems becomes more collaborated with artificial intelligence (AI) models for human comfort, quick and accurate predictions of individual thermal comfort from a human body will contribute substantially to optimize the heating adjustment. In this study, a methodological approach of accurate prediction, rather than just approximate estimation or assessment, of individual thermal comfort and sensation has been proposed especially for an AI driven heating adjustment. Called artificial intelligence vote (AIV), for realizing improved prediction accuracy, newly devised Individual thermal comfort indexes have been acquired from a linear regression machine learning (ML) model where thermal comfort is expressed as a combination of different categories of user-centered physiological variables together with PMV itself, followed by a design of experiment (DOE) analysis of AIV indexes for strengthening and hence ensuring the model soundness for further applications.</p>