Objective To explore the latent class and influencing factors between different behavioral lifestyles in community patients with diabetes, to classify the population and reveal the behavioral lifestyle characteristics of different types of diabetes, and to provide a scientific basis for active health management of diabetes, the identification of high-risk groups and maximization of the effect of intervention measures. Methods From June to August 2022, a behavioral and lifestyle follow-up survey was conducted on 1 179 diabetes patients with electronic health records from 18 community health service centers in Pingshan District, Shenzhen. The latent class analysis (LCA) method was used to cluster the life and behavior style of the study subjects, and the disordered multi-classification logistic regression was used to analyze the influencing factors of each class. Results The LCA results showed that there were three latent classes of community diabetes patients: ^“overweight and obese group^” (28.4%), ^“balanced diet group^” (30.8%), and ^“vegetarian but persistent exercise group^” (40.8%). Logistic regression analysis results showed that age, total cholesterol , LDL , fasting blood glucose, comorbidities, and hypertension history affected the latent classes of behavioral lifestyle in diabetic patients in the community (P<0.05). Conclusion The behavioral lifestyle of community diabetes patients has obvious classification characteristics and obvious differentiation in terms of eating behavior. Doctors and health managers can issue corresponding lifestyle medicine prescriptions and formulate corresponding active health management measures to prevent the occurrence of risk factors related to diabetes patients in advance and formulate personalized intervention measures.