本文提出一种基于多维定标的无线传感器网络三维定位算法,结合RSS经验衰减模型和最短路径建立相异性矩阵,采用轻量级矩阵分解算法降低相异性矩阵分解的计算复杂性,并利用网络中存在的周期性消息将初始定位信息回送,在后台使用迭代优化算法对初始定位结果求精。仿真实验表明,在测距误差一定的情况下,该算法能够提高节点三维坐标的初始计算精度,经过集中式的优化求精后与MDS-MAP算法相比,能够明显地提高节点三维定位的精度-This paper presents a method based on multidimensional scaling in wireless sensor network positioning algorithm, combined with RSS experience attenuation model and the shortest path to establish dissimilarity matrix, using lightweight matrix factorization algorithm reduces computational complexity dissimilarity matrix decomposition, and the use of the network in the presence of periodic messages will be the initial positioning information return, in the background using an iterative optimization the algorithm to the initial positioning results refinement. Simulation results show that, in some cases ranging error, this algorithm can improve the calculation accuracy of three-dimensional coordinates of the initial node, through the centralized optimization refinement after compared with MDS-MAP algorithm, which can obviously improve the precision of 3D Node Localization