Data mining technology is a commonly used method to induct and summarize the experience of famous doctors based on the constructed database of medical cases. To promote the use of the data mining results and to explore the method suitable for constructing the database of medical cases from modern famous veteran doctors of traditional Chinese medicine（TCM）

this study optimized the database structure and common fields

used HanLP word segmentation system in combination with maximum probability algorithm-based hidden Markov model（HMM） framework to perform text segmentation and semantic annotation

and extracted the data according to the annotated semantics through python technique. This method helps to optimize the data structure of medical records

and improves the efficiency of database construction and the quality of data mining.