Power security is a crucial guarantee for the sustainable development of modern society. With the gradual development of information technology, the increasing number and strength of cyber attack methods can cause severe damage to new power systems. Reasonable network topology and effective network defense resources are key to load recovery after a power system suffers a cyber attack. Therefore, a strategy for resilient topology optimization and defense resource allocation of the cyber-physical system (CPS) of distribution networks under FDIA-Worm hybrid attacks is proposed to enhance the resilience of distribution systems against cyber attacks. This model adopts a three-tier framework of upper, middle, and lower levels to optimize the topology and defense resources: the upper level establishes a multi-objective Pareto planning model with planning costs and load loss risks as objectives, combines it with the middle-level network attack propagation model that considers attacks and recovery, and uses the non-dominated sorting genetic algorithm II (NSGA-II) to solve the planning scheme; the lower level considers various coupling schemes between the information layer and the physical layer, and evaluates the optimal topology configuration based on resilience metrics of the CPS of distribution networks. Compared with traditional one-to-one series mode schemes, the optimized network topology and defense resource schemes under the three coupling relationships obtained through model solution can play a significant role in enhancing system resilience.