The invention discloses a multi-time-scale grid-source-load-storage coordinated control method comprising the following steps that (1) the adjustable resources existing in a power distribution network and the adjustment scope under different time scales are analyzed; (2) a day-ahead optimization model is established; (3) day-ahead optimization is performed on the distributed power supplies, energy storage and flexible load under the active power distribution network by using a particle swarm algorithm according to the day-ahead optimization model and day-ahead forecasting information so as to ensure the total operation cost to be minimal within the optimization period; (4) an intra-day optimization model is established; and (5) rolling correction is performed on the operation state of the distributed power supplies, energy storage and the flexible load and the operation mode of the power distribution network of the current period of time according to the power flow distribution situation of the current power distribution network, the node voltage situation, the operation state of the distributed energy and the charge and discharge situation of energy storage with combination of ultrashort-term load forecasting and power generation forecasting. The effect of multiple adjustable and controllable resources in operation of the active power distribution network can be fully exerted so as to guarantee effective consumption and efficient utilization of the distributed power supplies.