Node self-localization is one of the key technologies in wireless sensor network (WSN) applications. Aiming at the problem of inaccurate RSSI ranging process due to environmental interference, this paper presents an algorithm of weighted compensation of coordinate error after taking the environment similar feature into account between reference node and blind node. The algorithm first selects the three suitable reference nodes to calculate the preliminary coordinate of blind node by using the signal propagation loss model and the centroid of intersection model, then getting three groups coordinate errors by reference nodes location each other. Finally, as coordinate error weighted to the blind node coordinate, getting the ultimate estimate coordinates of blind node. This paper also makes some analysis on the choice of reference distance. Simulation results in MATLAB show that this algorithm can improve the positioning accuracy effectively，and choose a different reference distance which has different positioning accuracy.