A comprehensive study of the Dynamic Framed Slotted ALOHA(DFSA)-based anti-collision multi-tag identification algorithms for passive Radio Frequency Identification(RFID) system is performed, and the tag estimation techniques are sequential weighted. Performance of DFSA algorithms before and after the improvements is compared through extensive simulation and the optimal forgetting factor is determined. The improvements can shorten the time required for identifying tags to improve the work efficiency of the RFID system. It can be effectively applied to container management, logistics management of ports, railway stations or logistics center, and solve the problems of slow goods flow or congestion caused by the multi-tag collision in RFID systems.