Abstract: As wind power gradually replaces the traditional power sources, there has been a deterioration in the system frequency adjustment. The active participation of wind power in Load Frequency Control(LFC) of an interconnected system becomes a new way to improve the system frequency characteristics. In response to this background, based on the distributed Model Predictive Control, this paper proposes a multi-source collaborative LFC strategy for interconnected systems with high wind power permeability by overall considering the frequency modulation resources and their response characteristics of the traditional units, the wind turbines and the energy storage in an interconnected system. Firstly, by analyzing the impacts of different wind speeds on the frequency modulation characteristics of the wind turbines, a multi-wind-speed-stage power response model for the wind turbines is proposed; Secondly, the LFC model of the interconnection system is constructed involving the traditional units, the wind turbines and the energy storage power stations, and the frequency response constraints of each of the units are taken into account. The weighted function with the Area Control Error signal and the automatic generation cost of the interconnection system is taken as the objective to establish the distributed model prediction controllers of the areal information interaction; Finally, in order to achieve a global optimal control of the load frequency of the interconnected system, each controller combines the operating status of units in its own area with the other areas to solve the power reference values of all the units online. The simulation results show that the proposed strategy effectively reduces the amplitude of the system frequency and the tie line power fluctuations, achieving the optimal power allocation among the units and reducing the cost of automatic power generation of the system as well.