To address the problems of high computational complexity, inflexible frame length adjustment, and sub-optimal system efficiency of the RFID tag anti-collision algorithms in the Internet-of-Things systems, a low-complexity, and universal fast RFID tag anti-collision algorithm is proposed in this paper. A faster and less-complex tag number estimation method depends less on computing and storage resources, making it easier to integrate into the Internet of Things. Using the concepts of sub-frame and system efficiency priority, after each sub-frame is identified, the number of tags is estimated quickly and the frame length is adjusted dynamically to ensure the algorithm's efficiency. Moreover, the proposed algorithm is fully compatible with the EPC Class-1 Generation-2 standard, which ensures its universality and compatibility with the existing systems. The simulation results show that the proposed algorithm can achieve a system efficiency of 0.3554, a time efficiency of 0.7851, and an identification speed of 433 n/s. Compared with the standard Q algorithm, the performance is improved by 9.691%, 5.002%, and 8.250%, respectively. It is, hence, demonstrated that the proposed algorithm meets the requirements for the rapid identification of the RFID tags in the Internet-of-Things applications. INDEX TERMS Anti-collision algorithm, Internet of Things, low cost, RFID tag, system efficiency priority. I. INTRODUCTION Radio-frequency identification (RFID) technology has been widely used in industrial, agricultural, and commercial production systems such as warehouse management, logistics tracking and supply chains [1], [2], biological asset inventory, and ticketing systems. The integration of RFID and sensing technology makes it possible for RFID sensor tags to function as intelligent sensing nodes in the Internet of Things and has led to additional research results and commercial products [3]. Passive Ultra High Frequency (UHF) RFID technology is based on the EPC Global Class-1 Generation-2 (EPC C1G2) standard. UHF RFIDs have the advantages of fast identification, no need for a power supply, and a long communication distance. They have been widely studied and applied in Internet of Things systems [3], [4]. The associate editor coordinating the review of this manuscript and approving it for publication was Vyasa Sai.