Abstract: After analyzing many typical association rule mining algorithms and the maximal frequent item set mining algorithm,a new algorithm of maximal frequent item set mining,named as BOFPV_MMFIA,is proposed.A frequent item set vector FP-V is introduced so as to convert the course of maximal frequent item set mining into the course of AND operation of the FP_V vector.The existing Apriori and its related algorithms produce a lot of candidacy sets and need scanning database many times,and the BOM algorithm entails the AND operation of k vectors with times.Overcoming these drawbacks,the proposed BOFPV_MFIA algorithm needs scanning database only once.Therefore,the proposed algorithm is obviously superior to Apriori and BOM algorithm in efficiency.