In the testability verification test for electromechanical equipment,the confidence level is usually low since the sample allocation method is not rational.To solve the problem,we proposed a new sample allocation method based on multi-factor by analyzing the key factors that have effect on the confidence level of conclusion.The sample size of each Unit Under Test (UUT) was calculated out according to the relative ratio of integrated multi-factor,which could optimize the failure sampling structure and improve the representation of failure sampling set.Moreover,the single-step failure propagation algorithm was improved,and the improved algorithm can calculate failure propagation intensity more effectively.We made sample allocation by taking a certain stabilization and tracking platform as an example.The result showed that the scheme can optimize failure sampling structure and improve the confidence level of the testability verification test conclusion.