Objective To explore the influencing factors of social frailty of community-dwelling older adults and their network relationships, and to provide a theoretical basis for targeted intervention. Methods The general data questionnaire, the Help，Participation，Loneliness，Financial & Talk Scale （HALFT）, the Athens Insomnia Scale, the FRAlL Questionnaire, the Social Support Rating Scale, Mini Nutritional Assessment Short Form （MNA-SF）, the Mini Mental State Examination （MMSE) and the Rasch-derived short form of the Center for Epidemiologic Studies-Depression scale (CES-D) were used to investigate 1 055 elderly people in Lubei District of Tangshan city. Results The incidence of social frailty among the community-dwelling older adults was 23.70%. Logistic regression analysis showed that age, education level, number of chronic diseases, sleep status, social support, nutrition status, physical activity, physical frailty, cognitive function and depression were the main influencing factors of social frailty（all P<0.05）. The results of Bayesian network model showed that age, physical frailty, cognitive function, depression and social support were directly associated with social frailty; physical activity was indirectly associated with social frailty through the number of chronic diseases, nutritional status and depression; older adults ≥70 years of age, and those with cognitive impairment and low level of social support, physical frailty and depression had the highest probability of social frailty (79.0%). Conclusion It is common for the community-dwelling older adults to suffer from social frailty. Medical staff should pay attention to the high-risk groups of social frailty, and it is necessary to reduce the risk of social frailty through the collaboration of multidisciplinary teams and comprehensive intervention measures.