Abstract: In this paper, to enhance the task execution efficiency of multifunction radar with a partitioned aperture, a Double Deep Q-Networks (DDQN) based dynamic task scheduling method is proposed. Firstly, the basic task model and task scheduling model of multifunction radar are established according to the task characteristics of multifunction radar with a partitioned aperture. Then, by analyzing the radar multi-task state information and its change process, we construct the task scheduling process of multifunction radar as a Markov decision process (MDP), in which we design the state space, action space and reward function. Finally, considering the impact of current task scheduling decision on subsequent task scheduling decisions, a DDQN-based task scheduling method is formulated to realize efficient dynamic task scheduling for multifunction radar with a partitioned aperture. Simulation results show that compared with the traditional task scheduling method for multifunction radar with a partitioned aperture, the proposed method significantly reduces the task drop ratio of multifunction radar in a high load environment.