Objective To establish an individualized bleeding risk assessment system for the elderly coronary heart disease (CHD) patients complicated with intestinal malignant tumor based on single-center clinical big data. Methods Clinical data of the elderly CHD patients with intestinal cancer and being treated in the Chinese PLA General Hospital during January 2008 and December 2018 were collected retrospectively from the Clinical Database in the Big Data Center of the hospital, and they were subjected as the validation cohort. Taking the occurrence of major as the research endpoints, baseline analysis, decision tree model, support vector machine, logistic regression model and random forest model were performed on the clinical data. And then the CHD patients with intestinal tumor admitted into the hospital from January 2019 to December 2020 were prospectively recruited and served as derivation cohort. Finally, the performance of above models were evaluated and verified based on the accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). A predictive system for bleeding risk was established on the obtained optimal model. SPSS statistics 15.0 and R 3.6.1 were used for statistical analysis. Data comparison between two groups was performed using student′s t test, Chi-square test or Wilcoxon test depending on different data types. Results There were 511 patients in the derivation cohort and 35 patients with clinically significant bleeding events; 205 patients in the validation cohort and 11 patients with clinically significant bleeding events.Recursive feature elimination was used to screen the variables, and the logistic regression model containing 5 variables (brain natriuretic peptide precursor, total bilirubin, aspartate aminotransferase, carcinoembryonic antigen and urea) was selected as the optimal model. In the training set, the AUC value, accuracy, sensitivity, and specificity of the model were 0.791,0.757,0.714, and 0.800, respectively. In the verification set, the AUC value, accuracy, sensitivity, and specificity were 0.748,0.747,0.500 and 0.760, respectively. Based on this model, we constructed the bleeding prediction score for clinical application. Conclusion Our establised risk model and score system can predict bleeding events in the elderly CHD patients with intestinal malignant tumor.