In this paper, a cloud-edge-end collaboration-based control architecture is established for frequency regulation in interconnected power systems (IPS). A model predictive control (MPC)-based load frequency control strategy for the IPS with photovoltaic aggregation and energy storage systems under model uncertainty and communication delay is proposed. This can efectively overcome the issues of model uncertainty, random load perturbation and communication delay. First, a state space model for the IPS is constructed. To coordinate the frequency and contact line power fuctuation of the IPS, a robust controller based on the theory of MPC is then designed. Then, considering the communication delay of frequency response commands during transmission, a predictive compensation mechanism is introduced to eliminate the efect of delay while considering model uncertainty. Finally, simulation results verify the efectiveness and robustness of the proposed control strategy.