With the expansion of the distribution network scale, the traditional centralized state estimation algorithms require increasingly complex model construction, leading to increased computational complexity during the solution process, reduced timeliness and accuracy of the state estimation. Aiming at this problem, a distribution network state estimation method based on node degree search partitioning is proposed. Firstly, a node degree search partitioning method based on balanced regions is proposed, and the measurement model and distributed state estimation model for the distribution network are designed. Secondly, the distributed state estimation model is solved in three stages. Finally, the IEEE 30 node distribution system is selected for simulation analysis. While ensuring the observability of the system after partitioning, a comparison is made among different partitioning methods in terms of the accuracy and duration of state estimation. The impact of measurement redundancy on state estimation accuracy is further analyzed. The simulation results demonstrate that the proposed distributed state estimation method has significant advantages in terms of state estimation speed and accuracy.