Abstract: Frequent extreme natural disasters are highly likely to cause large-scale faults in the distribution network,and the distributed generator（DG） based load recovery is an important means to improve the resilience level of the distribution network.However,existing studies generally make independent decisions during the in-and post-disaster stages,while neglecting the continuity and coupling of the two decisions,and slowing down the recovery process.Therefore,this paper proposes a fast recovery method for the distribution network based on the in-and post-disaster two-stage coordinated decisions,which fully considers the interactions between the in-and post-disaster processes,and integrates the recovery measures including DG scheduling,network topology change,and maintenance personnel scheduling to guarantee the optimality of the consecutive decisions during the in-and post-disaster stages.Firstly,considering the demand relationship between post-disaster energy supply path and the in-disaster network topology,the topology adjustment model with DG as the core and the cellular load recovery model are established during the in-disaster stage to achieve the power supply for partial important loads.Secondly,considering the impact of the in-disaster decisions on the post-disaster recovery,a joint decision-making model for the fault maintenance personnel scheduling and fast load recovery is established during the post-disaster stage to achieve orderly recovery of all loads.Then,with the objective of minimizing the loss of load during the in-and post-disaster stages,the coordinated decision-making for the in-and post-disaster recovery is achieved.Finally,the validity of the proposed model is verified through an IEEE 123-bus distribution system.