Abstract: Aiming at the problem which positioning accuracy of unknown node overly depends on physical measurement accuracy and anchor nodes density among the localization algorithms based on RSSI in wireless sensor networks, an error self-calibration genetic localization algorithm based on RSSI is proposed in this paper. In the first period of positioning, in order to eliminate the measurement error, the maximum received signal strength indicator anchor node within communication range of the unknown node is selected as the bias reduction reference minutiae; In the second period of positioning, a modified fitness function is introduced, which weaken the phenomenon of unreal fitness. The simulation results show that the positioning accuracy of the proposed algorithm is higher than that of genetic localization algorithm under the same simulation condition.