The invention relates to a management optimization method in the field of alternating-current micro grids, and specifically to a multi-time-scale micro grid energy management optimization scheduling method. Aiming at the problems that the existing studies are incomprehensive on consideration of the running status of a micro grid, not detailed enough on the control strategy in each running status of the micro grid, insufficient on calculation accuracy of the running cost and the like, the method comprises a day-ahead economic optimization scheduling phase and an intraday economic optimization scheduling phase; the former considers the electricity price at peak, valley and flat time, and distributed unit power optimization allocation in the micro grid is performed according to day-ahead photovoltaic and load prediction by using the total running cost including the running and maintenance cost of a lithium battery and a fuel battery, interrupted compensation of an interruptible load, price of electricity sold by a large grid and the like as a target function; and the latter considers the outputs of electricity purchased from the large grid by the micro grid, a super-capacitor and the fuel battery, and an independent control strategy is respectively established for the peak, valley and flat time, so that the control strategies are more specific and detailed.