Existing business end programming methods lack guidance for end-users without development experience. So a business end programming approach based on Bayesian network was proposed. First, a basic framework based on Bayesian network for end-users programming was presented. Then the domain expert system which is used to support the business end programming approach was established based on Bayesian structure learning method and process logs. On this basis, a heuristic method for business end programming taking domain expert system as core component was proposed. This method could recommend business activities to end-users in the process of programming relying on domain expert systems. It also provided real-time guidance and help end-users to complete programming step by step. Finally, the efficiency of the method was evaluated and analyzed, and the results show that the business end programming method based on Bayesian network could improve the efficiency of end-users programming to a certain extent.