Abstract
The health of older adults is related to physical activity, and physical activity level (PAL) is associated with various factors such as community environment and psychological factors. Based on the above assumptions, this study constructs Structural Equation Modeling(SEM) to examined the relationship of psychological variables with older adults' PAL. The study further combines observable variables such as socioeconomic attributes, travel characteristics, and objective community environment elements with latent variables to construct SEM-Logit model to examined the relationship between these factors and PAL. Our results showed that: (1) PAL was related to gender, age, living conditions, self-rated health, PA scope; (2) PAL was positively related to physical activity intention (PAI), and other perceived psychological variables had an indirect effect on PAL through PAI; (3) Older adults living in a walkable community are more likely to improve their PAL. Corresponding strategies were proposed from improving the community environment and guiding psychological perception.