Abstract: The algorithm of discovering association rules will be introduced to data mining of a large database of Chinese medicine formulations in order to find the association rules among the unilateral and binding medicine of Chinese medicine to provide an important basis for the development of new drugs.Since the Apriori which is a kind of association rules algorithm that commonly used has the shortcoming of scanning the database for many times.This paper presents a matrix-based algorithm of association rules Apriori_Matrix.The algorithm shortens the time spent in several comparisons in collection connection in Apriori algorithm，and can greatly improve efficiency of the association rules dig.For the limitation of the type of the unilateral in the traditional Chinese medicine database and great difference of compatibility rules and a variety of prescriptions corresponding one type of disease，improving algorithm can shorten the development cycle of new drugs.