Abstract: Along with ever-increasing proportion of such loads as air conditioners and electrical vehicles in urban load, these kinds of loads play a part in the recovery process of post-fault urban power network. Thus, a curve describing the variation during the start-up process of the cooling load controlling indoor temperature, in which the influences of ambient temperature and outage duration was taken into account, was constructed, and according to different outage durations, ambient temperatures and power off times the multi-scenario simulation for load recovery processes was carried out. In the load recovery strategy, by means of network cohesion degree and nodal load amount the importance of load nodes was evaluated, and the priority of recovering load was considered. To improve the reasonability of power flow distribution during the load recovery process, the probability of appearing heavy-loaded transmission lines during the recovery process was reduced as far as possible, and the improved weighting power flow entropy was put forward and taken it as one of the objective functions for load recovery. Calculation results of IEEE 30-node system show that by means of the proposed model and strategy the safe and efficient recovery of post-fault urban power network can be implemented, so the enaction of power system recovery plan is more adaptible for the load variation of urban power network.