Dementia is a syndrome consisting of a group of symptoms that cause progressive brain dysfunction and impairment of cognition or intellects. Dementia causes problems with memory, orientation and thinking processes to the extent that it decreases functionality in Activities of Daily Living (ADL) and causes behavior change. Several studies have focused on the unusual behavior detection and pattern detection of daily activities using in-home sensors. Activities are detected either through specific rules or by applying learning techniques. This study focuses on utilizing adaptive algorithms where the relationship between environmental and cognitive state will be predicted. To automatically assess cognitive health based on activity and behavior, this research will focus on introducing a method to automatically predict the behavior of patients at homes in a smart way and predicting the patient’s cognitive state.