Abstract: Association rules discovery is an important research topic in data mining.The article will mine Data-Item set compression to a Boolean Vector Matrix.The algorithm only needs to scan the database one time and rational use of data storage structure so that it would not generate a large number of candidate sets.Experiment results indicate that the algorithm is not only simple，but also has good efficiency compared with the Apriori algorithm.When you mine a large database and long item-set，the mining effect of this way is more visible than the Apriori algorithm.