In order to enhance the resilience of the distribution network and cope with the development trend of new power grids, this paper proposes a particle swarm algorithm based on the goal-oriented adaptive for siting and capacity setting of distributed power supply, so as to explore the dynamic reconfiguration optimization problem of the distribution network containing distributed power supply. The method firstly improves the inertia weight factor by using the goal-oriented adaptive method to improve the algorithm's ability of iterative optimization search, and calculates the optimal access location and capacity allocation of distributed power supply. Secondly, the power output of distributed power supply is analyzed by using the K-means and ISODATA clustering methods, and the DBI and DI indices of the two methods are calculated to analyze the clustering effect. Finally, the dynamic reconfiguration analysis of multi-fault distribution network is considered with voltage loss and load abandonment as the objective function, and overhaul strategy with constraints of manpower and material resources as the constraints. The analysis of the results shows that the introduction of the goal-oriented adaptive method improves the convergence speed of the particle swarm algorithm; the distributed power supply situation is analyzed by the ISODATA clustering method, and the clustering effect is better; finally, the dynamic reconfiguration study of the distribution network is carried out by using the mixed-integer programming method to improve the reliability of the distribution network.