Abstract: Objective To construct a depression risk prediction model for elderly people with cardiovascular disease（CVD） in China,providing a basis for early identification of depression. Methods The empirical analysis was conducted by using the fourth round of China health and retirement longitudinal study（CHARLS） survey project,which was officially released to the world on September 23,2020. The center for epidemiological studies depression scale（CES-D） was used to evaluate the depressive mood of elderly people with CVD. The non-parametric tests and chi square tests were used for univariate analysis,and multivariate logistic regression was used to analyze the influencing factors of depression in elderly people with CVD. A depression risk prediction model for elderly people with CVD was constructed,and it was randomly divided into training and validation sets in a 7∶3 ratio. The model was calibrated to verify its effectiveness. Results A total of 4 325 respondents were included. Women,decreased educational level,separated/divorced/widowed/unmarried,decreased self-reported health,disabilities,chronic lung diseases such as chronic bronchitis or emphysema（excluding tumors or cancer）,arthritis or rheumatism,kidney diseases（excluding tumors or cancer）,stomach or other digestive system diseases（excluding tumors or cancer）,and engaging in vigorous physical activity were influencing factors for depression in elderly CVD patients（all P<0.05）. The ROC curve AUC sizes for constructing the training and validation sets of risk prediction models were 0.734（95%CI：0.717-0.752） and 0.749（95%CI：0.722-0.776） respectively,with Hosmer Lemeshow test showing P>0.05. The two groups have good goodness of fit,and the risk prediction model has good calibration. Conclusions Women,decreased education level,separation/divorce/widowhood/unmarried,self-rated health decline,disability,chronic lung diseases such as chronic bronchitis or emphysema（excluding tumors or cancer）,arthritis or rheumatism,kidney diseases（excluding tumors or cancer）,stomach or other digestive system diseases（excluding tumors or cancer）,and intense physical activity significantly increase the risk of depression in CVD elderly people. After a series of validations,it is suggested that both the training and validation sets of this model have a net benefit range,good consistency and predictive performance,providing a basis for early screening of depression in CVD elderly people by medical staff.

Key words: Cardiovascular disease, Depression, Risk prediction model