The satellite scheduling schemes in practical applications are typically affected by external factors, such as electromagnetic interference, satellite failure, cloud cover, and the dynamic arrival of new tasks, etc. It is necessary to adjust the original scheduling scheme within a short time and generate a new scheduling scheme to ensure the stability of the satellite system. Therefore, considering satellite emergency scheduling when new tasks arrived, this paper establishes an integer programming model for multi-satellite cooperative observation emergency scheduling, and proposes a satellite emergency scheduling algorithm based on a dynamic neighborhood(SESA-DN). This method designs various types of neighborhood structures and dynamic neighborhood selection strategies, which can provide effective feedback in terms of the completion of tasks, and generate satellite emergency scheduling schemes by inserting and replacing emergency tasks. Besides, a multi-satellite emergency scheduling scenario is constructed, and the proposed SESA-DN algorithm is compared with multiple comparison algorithms through a large number of simulation experiments. Experimental results show that the SESA-DN algorithm is more stable than comparison algorithms, and has good applicability to the multi-satellite cooperative dynamic scheduling problem.