Due to the lack of valid testability evaluation methods for complex equipment in the final phase, a new testability evaluation method based on test data in development phase is proposed. The problems that classic statistic method can produce the conclusion with low confidence level and high risk under the condition of small sample are effectively solved. This method adopts mixed prior distribution and determines the inheritance factor by using chi-square goodness of fit between the history test information and the field information. Finally, the fault detection rate of certain large-scale radar equipment is validated. The analysis result of the practical case proves that this method can produce evaluation conclusion with high confidence level, and shows that this method is more rational than the classical statistical method and the traditional Bayes method.