According to the uncertainty existing in the process of acquiring information of spatial ballistic target and its accompanying decoy, subjective Bayesian method was used to recognize these spatial targets. In order to satisfy the premise of necessary knowledge and experience for applying subjective Bayesian method, neural network was used to replace the experts of relevant domains. The method to combine neural network and subjective Bayesian was introduced, and a multi sensor data fusion model to recognize spatial complex ballistic taget and its accompanying decoy was given.The results of emulation show that the recognition rate is improved much.