Abstract:As medical records of traditional Chinese medicine (TCM) carry academic thought and clinical experience of famous TCM doctors, their researches are important think-tank resources to promote the discipline development of TCM, and effectively improve the clinical efficacy. With the development of information science and technology, in the face of massive data in medical records of TCM, we shall dig out hidden knowledge, thought connotation and medication regularity by using the data. Through the study of medical records, we can constantly summarize previous experience and explore academic thought. The most traditional method for analyzing TCM records data is the comprehension-based analysis method. However, due to differences in individual level and research purpose, the conclusions were highly subjective and time-consuming, with a poor promotion. The general statistical method is to establish database, or make simple analysis of symptoms and drug frequency. To quickly find academic thoughts and clinical experience of physicians from massive medical records, we need to introduce new technologies and methods, and integrate them with databases, artificial intelligence, statistics, knowledge engineering, information retrieval, high-performance computer and data visualization. From the perspective of research method, this paper introduces the personal comprehensive-based analysis method, frequency analysis method, multivariate (factor) analysis and other research methods, so as to study the application of data mining in TCM researches, and provide research methods and ideas for further summarizing and inheriting clinical experience of famous TCM doctors. The laws hidden in medical records of TCM were dug out, and will applied in scientific and technological researches and clinical researches, in order to reflect the research value of medical records of TCM.