Based on the introduction of the concepts of several clustering algorithms, the paper puts forward the optimized K-means algorithm, applies it to the analysis on 2 000 pieces of clinical data about Traditional Chinese Medicine(TCM) treatment of high blood pressure, and gets 8 corresponding drug combinations of different symptoms. Through the result, on the premise of guaranteeing clustering quality, the optimized K-means algorithm improves the arithmetic speed of the algorithm, and is provided with obvious superiority of performance.