Advances in medical technologies and health sciences and population aging year by year have urged the development of smart healthcare. To cope with the shortage of caregivers for the rapid growth of the elderly population, this study installed the ambient sensors in the houses of the elderly to collect behavioral sensing data for their lifestyle analyses. This research applied machine learning to analyze the sensory data and compare them with the daily-activity records of the subjects. The differences of the lifestyles and behaviors between the subjects and their family members were also investigated. The results show that the proposed methods in this study have great potential to identify group-lifestyle differences and individual behavioral variation.