The article integrates the theoretical study, simulation validation and performance analysis to make a deep research on distributed implementation of iterative localization technology in wireless sensor networks. Firstly, we utilize Kalman filter method based on square-root cubature to estimate and correct the node’s position in real time. We establish a distributed iterative localization algorithm where the nodes that get localized in the current generation serve as references for remaining nodes to localize and the localization process is repeated. In order to improve the deficiency on location accuracy generated from propagating localization errors, and then we formulate the error feedback control method with Taylor series expansion as condition of evaluating whether a node succeed in localization, which is applied to iterative localization to establish a Taylor feedback Kalman filter localization algorithm. The simulation validation shows that the location accuracy of this kind of algorithm can fully meet the location requirement of the wireless sensor networks. Compared with the traditional method, the results illustrate the performance advantages of the error feedback control method and its contribution to the accuracy of the node position estimation.