Focusing on the problems of poor locating performance and ignoring anchor nodes' inaccuracy in traditional DV-Hop algorithm which is one of the range-free algorithms in Wireless Sensor Network (WSN), an intelligent algorithm for locating nodes based on Processing Strategy of Optimal Hopping Distances (PSOHD) was proposed. Fully considering the effect of dynamic topology and anchor nodes' difference, this algorithm firstly introduced two communication radii with anchor nodes to calculate the nodes of communication range respectively. Secondly, the average one-hop distance among anchor nodes was refined by means of weighted least squares estimation. Finally, the average one-hop distance used by each locating node for estimating its location was optimized through weighting the N received average one-hop distances from anchor nodes. In addition, Particle Swarm Optimization (PSO) algorithm was presented to locate the unknown nodes. The simulation results show that the improved algorithm has obviously better location precision at the cost of increasing appropriate energy. Thus, it is a practical scheme for WSN with both inaccuracy of anchor nodes and random dynamic topology.