The design of Course of Action (COA) is an important part for establishing an efficient network command and control system under network conditions, and is of great significance for efficient operation.To solve the problem of finding the optimum COA in a dynamic battlefield environment, a mathematical model of COA was built up by combining Dynamic Bayesian Networks (DBNs) with Learn Genetic Algorithm (LGA) under the known priori information.The prior knowledge was combined with Conditional Probability Table (CPT) to search the optimum COA through LGA.The simulation results verified the feasibility and effectiveness of the DBN based COA model with prior knowledge and the solution method of the model.