Abstract: In indoor RSSI localization, precision is low and dynamic tracing of variation in parameters cannot be achieved. Therefore, an improved gravitational search algorithm (GSA) is used for RSSI localization. Obtaining the parameters of the model through maximum likelihood estimation and using the least squares estimate as preliminary results, the GSA is then utilized to optimize the preliminary results and parameters. The algorithm has fast convergence rate, high accuracy, etc. The test results show that the algorithm can not only improve the location accuracy but also dynamically trace the variations in the RSSI-positioning mathematical model's parameters so that the model can adapt to the changes within the environment.