Abstract: In order to solve the problems of unreasonable selection and low screening efficiency of the existing soil and groundwater pollution risk control and remediation schemes in China, the site pollution risk control and remediation scheme recommendation system was constructed by using the structured hierarchical storage and search technology as well as case-based reasoning(CBR) and machine learning methods, which realized fast search to find matching source cases. By studying the realization way and content of case base, the structure design and system development of the scheme recommendation system were carried out, and the case retrieval query page based on Web technology was established. Similar cases were retrieved in the case database by using K-nearest neighbor algorithm and analytic hierarchy process, and the scheme recommendation function was realized, in order to provide reference for decision-makers to select relatively optimized site pollution control scheme, and improve the accuracy, intelligence, high efficiency and low cost of site pollution risk management in China.