基于数理统计理论中的中心极限定理，提出一种适用于系统可靠性评定的渐近正态法。该方法是通过将子系统或部件试验数据进行有效的分析处理，使之服从正态分布规律，从而由标准正态分布给出在不同置信度条件下的系统可靠性下限。计算方法简单，具有工程应用价值。 Based on the central limit theorem in mathematical statistics theory, an asymptotic normal method suitable for system reliability assessment is proposed. In this method, the test data of subsystems or components are effectively analyzed and processed to obey the normal distribution law, and the lower limit of system reliability under different confidence levels is given by the standard normal distribution. Calculation method is simple, with engineering application value.