Abstract: The incremental association rules mining is a hot point for data mining research in recent years,and the appearance of CAN-tree is an important milepost in development of incremental mining association.Although the CAN-tree is suitable for incremental association rules mining,the CAN-tree needs long construction time,then causes low mining efficiency;In order to solve this problem,a fast construction algorithm based on CAN-tree is presented,the algorithm adds the auxiliary storage structure based on Hash which can be used to reduce project-search time,so that the construction of CAN-tree speed can be improved;Then the algorithm adds leaves address sheet and modifies node structure and changes child pointer to parent pointer,which reduce the generation time of conditional pattern base,so that the efficiency of data mining can be improved.experiment results show that the algorithm is effective.