Abstract: In this paper, a hybrid positioning algorithm based on optimization is proposed for applications of wireless sensor networks (WSNs). The presented method starts from the initial node which has the most neighbors in the network. Then, the relative coordinates of the initial node and its neighbors are estimated by multi-dimensional scaling (MDS) algorithm and broadcasted. The un-localized nodes estimate their coordinates with the maximum likelihood algorithm based on their received coordinates and ranged distances to each neighbor. In the end, all nodes' absolute coordinates are obtained by coordinate translation. Furthermore, the global and distributed optimization strategies are integrated into the presented algorithm for improving the positioning precision. The simulation results show that the proposed algorithms are valid and can localize sensor nodes precisely.