The extreme disasters have brought severe shock to the safe operation of power system, and caused great damage to distribution power network equipment and large-scale outage. In order to accelerate the process of post-disaster recovery, quick repairing to the damaged equipment is required. However, owing to the uncertainty of the fault location and lack of interaction in repairing dispatching scheme, there exists the problems of mispairing and idleness of repairing crew, leading to the low efficiency of fault repairing and post-disaster recovery. This paper focuses on this problem and proposes a dynamic post-disaster repairing strategy based on rolling optimization. Considering complex fault scenarios, the outage nodes near normal-operating upstream nodes will be first repaired. The single-phase optimization model has been formulated to solve dispatching problem between repairing crews and certain fault points, in order to lower the weighted load loss. Furthermore, considering the repairing completion, power recovery and fault point detection, a rolling optimization decision process has been designed and repairing crews can be dispatched dynamically. The cases validate that the strategy proposed in this paper can effectively promote the efficiency of repairing, avoid the phenomena of mispairing and idleness of repairing crews and accelerate the process of post-disaster recovery.