Aimed at the problem that the SBN network model fails to meet the requirements of the dynamic performance and regularly and accidentally mistake target recognition, a new target recognition model is designed based on variable structure dynamic Bayesian network to improve the capability of target recognition under high dynamic and complex electromagnetic conditions of environment. This modified model is developed by Static Bayesian Network model, has a good dynamic expression and filtering function, makes up for the lack of SBN, and has a good fault tolerance capability. The simulation results show that the effect of target recognition based on dynamic Bayesian networks is better than that of target fused recognition based on parameter learning Bayesian. The accuracy of target identification and the stability of the algorithm are significantly improved. By so doing, the model effectively solves the problem of missing data and information in the process of target identification.