The deep application of artificial intelligence （AI） technology in the military field means that scientifically assessing the vulnerabilities caused by defects in AI technology to intelligent weapon systems has become an important issue affecting the development of intelligent weapon systems. This study uses an in-depth analysis of the vulnerabilities of intelligent weapon systems to construct 14 vulnerability assessment indicators. The combination weighting method， which integrates the analytic hierarchy process （AHP） and entropy method， is used to calculate the indicator weights. The cloud model evaluation method transforms the ambiguity and randomness in the evaluation of qualitative indicators into quantifiable cloud characteristics. The assessment conclusions are drawn by observing and calculating the similarity between the comprehensive assessment cloud and each standard cloud. Taking the vulnerability assessment of a certain unmanned aerial vehicle （UAV） system as a case study， the feasibility of the proposed method is demonstrated， exploring new approaches for evaluating the vulnerability of intelligent weapon systems.