Abstract: We propose an agent-based model of smart home with intelligent heating system, i.e., a kind of computational laboratory for experimenting with smart solutions and ambient technologies. The model consists of four components simulating behavior of inhabitants, weather changes impacting the use of the heating system, heat transfer of the building, three control logics of the heating system which are non-intelligent thermostatic system, learning home a predictive system operating with the room occupancy matrices and attentive home-an adaptive predictive system which builds its own representation of the presence of individuals in rooms. The model is used to evaluate different scenarios of living in smart home with respect to two measures (user discomfort, heating costs). The model was implemented in NetLogo with IODA extension.