Abstract: In the state estimation of distribution systems, it is difficult to fully monitor numerous network devices, resulting in insufficient measurement data. Besides, frequent topology switching urgently requires efficient algorithms. Therefore, this paper proposed an effective weighted minimum absolute value state estimation method for distribution network topology identification. Firstly, the traditional weighted minimum absolute value state estimation method based on linear programming was improved to improve the calculation efficiency. Secondly, a weighted minimum absolute value state estimation method based on mixed integer linear programming was formed by adding supplementary variables and constraints to the reformulated weighted minimum absolute value state estimation method to adapt to the topology identification problem. The simulation results on four example systems validate the excellent performance of the proposed topology identification method in different amounts of real-time measurements, high pseudo measurement errors, measurements corrupted by bad data, and unknown branch state scenarios.