Abstract: Distributed Photovoltaic (PV) is ever increasing in the power grid, as the current PV generation rapidly developed. However, it is difficult to directly coordinate the PV into the conventional power grid, due mainly to the intermittent and uncertain nature of PV power generation. As such, there is a great impact on the power flow of a system, particularly the volatility and uncertainty of the output and load demands of Distributed Generation (DG) power in grid-connected microgrids. A friendly way is widely expected that the PV can access the distribution network in the form of a microgrid for the enhanced DG absorption capacity. It is highly urgent to reduce the impact of such volatility and randomness on the energy transmission between microgrids and distribution networks. In this study, a multi-time scale coordinated optimization was performed on energy scheduling strategies using a Variational Modal Decommission-Model Predictive Control (VMD-MPC). Specifically, an MPC was a sort of optimal control with a closed-loop over a finite time domain, suitable for the nonlinear, time-varying, and uncertainty of the system. There was no differentiation scheduling on the forecast of PV power for each Controllable Micropower (CMS)in the microgrid operation because the load was directly applied in the previous multiple-time scale scheduling using MPC optimization. Consequently, some CMS (such as a Lead-Acid Battery, LAB) was run in a short time, high strength, and energy scheduling, whereas, some CMS (such as a Super Capacitor, SC) was only for a long and slow energy scheduling. Thus, the operating characteristics of CMS in different time scales should be considered in the optimization of scheduling. A VMD was utilized to acquire the different loads and subsequence in the PV series of frequency scales, thereby achieving the multiple coordinated optimization scheduling CMS models in different time scales. The scheduling model included a longtime scale of 1 hour and 15 min interval time for a short scale. Dispatching LAB, Micro gas Turbine (MT), and small Biomass Generator (BPG) were usually responded to the signals of a long-time scale. Scheduling SC, MT and BPG were responded to the signals of a short-time scale. The final scheduling was achieved for each CMS to realize the differentiated optimal processing of signals on different time scales, where the calculated values of each model were summed up. Then a feedback correction model was constructed to form a closed-loop control using the day-ahead scheduling, where the difference between the ultra-short-term forecast within the day and the day-ahead forecast was taken as the disturbance input, while the current operating state of the system was taken as the parameter, and the power increment of each CMS was taken as the control variable. The feedback correction effectively enhanced the robustness, while reduced the impact of the system that resulted from the uncertainty of load and PV output. As such, the optimal energy scheduling strategy effectively coordinated the grid-connected microgrids with multiple micro power sources and time scales. Taking a PV microgrid in North China as an example, an MATLAB software was used to simulate and verify the model, indicating optimal scheduling. Better feasibility, effectiveness, and economy of strategy were achieved from the perspectives of power scheduling and operating cost, compared with the traditional active power and MPC multi-time scale scheduling strategy. Accordingly, this finding can provide a practical and effective technical approach for high-permeability microgrids in energy trading under the environment of multiple renewable energy consumption and electricity market.