Abstract: With the increasing penetration of uncontrollable distributed generators such as distributed wind generator and photovoltaic generation,the uncertainty of distributed generator has become a key factor in distribution network reconfiguration.Hence,a stochastic network reconfiguration model was used to minimize the mathematical expectation of overall power loss,considering the constraints of power flow,voltage and topology.The stochastic power flow based on Latin Hypercube sampling-Monte Carlo simulation（LHS-MCS）was used to examine the chance constraints of nodal voltage and branch power flow.To improve computational efficiency,parallel undirected spanning tree-based genetic algorithm（PSTGA）was proposed to solve the network reconfiguration in parallel.The results of test in IEEE 33-bus distribution system have confirmed the rationality of model.The efficiency of PSTGA has been proved by comparing with undirected spanning treebased genetic algorithm,particle swarm optimization,ant colony searching optimization and improved harmony search algorithm.