Abstrat Indoor location information service has been the focus of people's research. However, the positioning effect is still not very satisfactory, which need more references. RFID poisoning is simple and low power, it is one of the most used indoor poisoning methods. However, it is very difficult to build the mapping model of “The tag received signal strength indicator (RSSI) and the distance (DIST) between tag and reader” in the complex indoor environment. Therefore, this study uses a particle swarm optimization (PSO)-based back propagation (BP) neural network (PSO-BP) to determine the relationship between the RFID signals and the position of a tag for an RFID-based positioning system. In addition, in order to improve the quality of training samples, the experimental data are pre-processed by Gauss filtering method. The detailed computational and experiment results indicate that the proposed PSO-BP can provide better predictions (including the accuracy, stability and convergence speed) than a traditional BP and GA-based BP neural network.