Abstract: The anti-aircraft artillery kill probability is mainly tested by indirect test method, and the direct test method with the highest credibility is studied. For the direct test, the definition of kill probability is determined, various test methods are analyzed, the direct test method with the quantity of killed targets as the criterion should be selected, and a detectable definition of the kill event is presented. The test design for judging kill probability by quantity of killed targets is made. Instead of finding the results in a table, the recursive calculation model is derived to achieve higher calculation accuracy. It is proved that the maximum likelihood estimation of kill probability cannot be simply used as the criterion. The relationship among indicator, test risk and the minimum kill frequency is analyzed. The total number of tests and the minimum quantity of targets to be destroyed are calculated for a weapon system. The results show that the direct test method is feasible and can support the optimization of test design. Key

Key words: anti-aircraftartillery, killprobability, testdesign, directtest, testrisk