Data mining, which aims at extracting interesting i nformation from large collections of data, has been widely used as an effective decision making tool. This paper br oke through traditional therapeutic evaluation meth od with only by individual experience and proposed an interactiv e data mining framework, by which therapeutic evalu ation will be automatically implement to maximize the adaptive capacity of data mining. During the processing, a series of data mining algorithms be applying and domain exper t may affect the mining behavior as a object factor s and how they affects the behavior is also discussed. The ex periment results demonstrate that the proposed a se ries of algorithms enable the Therapeutic Evaluation system robust and efficient. It is also observed that med ical application of the model in therapeutic evaluation offers the advantages of data mining and there is a n emerging need for mining medical application with data minin g techniques.