In order to solve the problems in the existing frequent pattern mining algorithms, such as complex tree building and low mining efficiency, a high-performance algorithm for mining frequent patterns was proposed, namely B-List Frequent Pattern Mining (BLFPM). A new data structure of Building list (B-list) was employed to represent frequent itemsets, and the frequent patterns were directly discovered by intersecting two B-lists without scanning the database. Aiming at the high time complexity of connecting two B-lists, a linear time complexity connection method was proposed to improve the time efficiency of BLFPM. Besides, set-enumeration search tree and an efficient pruning strategy were adopted to greatly reduce the search space and speed up the execution. Compared to N-list and Subsume-based algorithm for mining Frequent Itemsets (NSFI) and prepost algorithm, the time efficiency of BLFPM was improved by about 12% to 29%, and the space efficiency of BLFPM was improved by about 10% to 24%. The experimental results show that BLFPM has good performance for both sparse database and intensive database.