In order to solve the problem of distribution network reconfiguration with Distributed Generation (DG), a reconfiguration model aiming at system loss, load balancing and voltage deviation is established, and the improved firefly algorithm is used to optimize the calculation. In view of the defects of traditional firefly algorithm, such as premature convergence and over-reliance on control parameters, the inertia weight is introduced and chaotic theory is adopted to adjust the parameters of the algorithm so that the algorithm balances the global search and local search capabilities. At the same time, the elitist strategy is introduced into firefly algorithm which speeds up the algorithm convergence. In order to reduce a large number of non-feasible solutions in network reconfiguration, by simplifying the network topology, the coding dimension is shortened and the efficiency of optimization is improved. Finally, some examples are given to verify the effectiveness and practicability of the proposed algorithm. This work is supported by National Natural Science Foundation of China (No. 61703242).