In radio frequency identification (RFID) systems, the rapid identification of all tags located within the range of a reader is a major research topic. For rapid identification, the development of an effective anticollision protocol between the reader and tags is essential. Numerous anticollision protocols have been proposed, but these anticollision algorithms tend to be used in a static environment where the tags available within the range of a reader do not change during an identification process. By contrast, in the dynamic environment of RFID systems, such as in inventory management, some identified tags are removed and new tags are included. The reader in such an environment must regularly reidentify the tags. The previously proposed dynamic identification method cannot fully utilize the results from a previous identification process, and that results in collisions and idles occurring again in the new identification process. Therefore, such an identification process is inefficient. In this study, an efficient protocol termed the leaf node bit collision detection-based query tree (LBQT) algorithm is proposed. The proposed LBQT algorithm can reduce unnecessary queries for unchanged tags and rapidly identify the new tags. Experimental results indicate that the proposed algorithm outperforms previous protocols without depending on various densities of tag distributions.