针对载人飞船飞行试验过程中故障理论与决策特点和可能应用的需要,介绍了神经网络技术的发展,分析了神经网络用于故障诊断的典型模型的优缺点,并提出在载人飞船故障诊断和决策中应用神经网络技术的一个系统模型框架,该框架对于建立飞船工程指挥辅助决策系统具有参考价值. In view of the need of the theory and decision-making of fault theory and possible application in the flight test of manned spacecraft, the development of neural network technology is introduced. The advantages and disadvantages of the typical model of neural network used for fault diagnosis are analyzed. And a system model framework for applying neural network technology in decision-making. This framework has reference value for establishing a command and decision support system for spacecraft engineering.