Due to the inverse ratio between prediction accuracy and time span, an active distribution network multi time scale economic dispatch optimization model was established, taking into account the errors in load prediction and distributed generation (DG) output prediction. The safety and economic operation of the power system can be guaranteed through three time scale rolling correction methods: daily ahead planning, intraday rolling optimization, and intraday real-time optimization. The experiments show that the multi time scale economic dispatching decision-making optimization model for active distribution networks can effectively ensure that the voltage at each node of the power system does not exceed the limit by adjusting the output of DG, capacitor compensation capacity, and the tap points of the OLTC, while promoting the absorption of DG and reducing unbalanced power. That is, it ensures the voltage quality and safe operation of the power system, and improves the economic benefits of the power system.