The morbidity and mortality of heart disease are increasing in middle-aged and elderly people in China. It is necessary to explore relationships and interactive associations between heart disease and its risk factors in order to prevent heart disease. To establish a Bayesian network model of heart disease and its influencing factors in middle-aged and elderly people in China, and explore the applicability of the elite-based structure learner using genetic algorithm based on ensemble learning (EN-ESL-GA) algorithm in etiology analysis and disease prediction. Based on the 2013 national tracking survey data from China Health and Retirement Longitudinal Study (CHARLS) database, EN-ESL-GA algorithm was used to learn the Bayesian network structure. Then we input the data and the learned network structure into the Netica software for parameter learning and inference analysis. The Bayesian network model based on the EN-ESL-GAalgorithm can effectively excavate the complex network relationships and interactive associations between heart disease and its risk factors in middle-aged and elderly people in China. The Bayesian network model based on the EN-ESL-GA algorithm has good applicability and application prospect in the prediction of diseases prevalence risk.