The invention relates to an obstetrical special disease association index mining method based on an electronic health record. The method comprises the steps of constructing a basic information database, performing disease classification data association, performing data preprocessing, performing sampling to obtain a sample set, constructing a sub-feature screener, constructing a combined screenerand performing association index inspection. The method has the advantages that on the basis of mass electronic health record data, association index mining is carried out, and resource input is low;2, by means of a machine learning method, high-dimensional indexes are automatically analyzed, and a more comprehensive associated index set can be quickly mined; 3, a process-based association indexmining system can be established for the obstetrical special diseases, and the system is easy to popularize in different medical institutions. Therefore, the method provided by the invention is a powerful supplement of a traditional method, on one hand, researchers can be assisted in knowledge discovery, and time and manpower are saved, and on the other hand, a basis can be provided for early prevention and treatment of obstetrical diseases.