In this paper a voting algorithm is presented to mine the optimal ingredient pattern of Chinese medicine. First this paper visualizes the data about the curative effect of different prescriptions of Chinese medicine ingredients on several cardio-indexes. Then the features about the curative effect from the curves generated by visualizing the data are extracted. Finally, a voting algorithm is adopted in which votes are held in three stages: the preliminary vote generates the features of each test group and cardio-index by each sample; the metaphase vote obtains the value of the curative effect by each feature of the given group and index; the final vote mines the optimal ingredient pattern. Experimental results show that the method is effective for the kind of problems and instructive to clinic medicine experiment. Its potential applications include the development of new medicine, optimal combination investment of venture capital and gene series analysis etc.