A novel satellite scheduling framework based on the divide and conquer principle was proposed. Under this framework, an ant colony optimization algorithm was employed to distribute observation tasks to different satellite orbits. Then, an adaptive simulated annealing algorithm was designed to solve the satellite observation problem involved in each orbit. According to the feedback on the scheduling results at each orbit, the task distribution schema was adjusted. This process was repeated until the termination condition was met. To improve the efficiency of the algorithm, the domain knowledge of the satellite scheduling problem was considered into the heuristic information model of the ant colony optimization algorithm. Next, two neighborhood structures were designed in the simulated annealing algorithm. In addition, the dynamic selection strategy was used to choose the most appropriate neighborhood search structure. Extensive experiments show that the proposed method can reduce the problem complexity effectively, especially in solving the large-scale satellite observation scheduling problems, which exhibits extraordinary performance.