Data mining methods and applications have been widely studied in the medical fields.As a distinguished traditional medicine with distinct theoretical framework, Traditional Chinese medicine (TCM) has the clinical data as its core knowledge source to promote the clinical abilities and theoretical knowledge refinement.Generally, manual induction of the empirical knowledge from the daily clinical practices is one of the most important approaches for TCM knowledge distilling.Therefore, data mining methods will be a natural choice to efficiently promote the empirical knowledge distilling procedure.In recent years, the researches on the knowledge discovery from the TCM clinical data utilizing data mining methods are actively conducted.However, TCM mainly concerns and utilizes the terminological information in the clinical data, which are complex, heterogeneous and with multi-granular information elements.Also TCM may have the different evaluation criterion on the data mining results, which would be interesting and useful for modern biomedical sciences.Therefore, we introduce the main research topics of TCM clinical data mining and present the related experiences.To help the data mining researchers dig the diamond out of the large-scale TCM clinical data,we have a discussion of the distinct challenges and research issues that should be tackled in this promising field.We argue that the roadmap of TCM research should be a bedside to bench mode, in which the scientific discoveries from the real-world clinical practices will be the main impetus for TCM development.We consider data mining as the core computing approach to promote the realworld clinical scientific discoveries in the framework of the effectivenssdriven approach for the TCM researches.