In order to realize the precise control of power grid regulation strategy with minimum economic cost, a life cycle management technology based on multi-time scale analysis is proposed. By integrating short-term, medium-term and long-term regulatory behavior data, a behavioral prediction model and optimization framework are constructed. Simulation experiments and example analysis are carried out to verify the effectiveness of the proposed model and control strategy. The average accuracy of PV and wind power prediction is 97.53% and 96.27% respectively, the total daily operation cost of the system is 1 386.32 yuan, and the grid control strategy can adapt to the load demand changes and allocate power reasonably; The relative prediction errors of the three regulation stages decrease gradually, and the minimum relative error of the real-time regulation stage is 0.26%. The results show that the multi-time scale control technology based on life cycle management can significantly improve the planning and accuracy of the control behavior, and the application of this technology can not only improve the control ability of the power grid, but also provide stable support for the power security in the field of military industry to ensure the energy supply security of key facilities.