As a process of real-time state change, the battlefield situation information of air combat also has the characteristics of fuzziness, incompleteness and randomness. Aiming at the problems that the traditional weighting method is too one-sided, simple linear combination is not strong in explanation, and it is difficult for decision makers to accurately and objectively describe fuzzy information, and take into account the randomness of the battlefield for reasoning analysis, this paper proposes to use the improved multi-condition cloud model reasoner to evaluate the threat index of the target. On the basis of the proposed combination weighting method of objective entropy weight to modify subjective weight, the multi-conditional reasoners are divided into several single conditional reasoners and the rules are fused according to the weight, and the reasoning is repeated many times to reduce the error caused by randomness in the process of cloud reasoning. Finally, the threat level of enemy aircraft is obtained, and the specific threat index of enemy aircraft is also obtained. It provides a reasonable basis for the subsequent fire distribution between the enemy and us.