模糊多目标决策方法是并行工程初期产品总体设计方案选择的一种有效方法,但该方法必须进行一系列繁杂的推导过程。而在此提出的神经网络学习方法,只要对已有方案的模糊决策结果进行学习,就能得到同类产品其它方案满足于总体设计目标准确的隶属度值。通过YP—CIMS开发早期的多目标决策,证实了该方法的准确性和有效性。 Fuzzy multi-objective decision-making method is an effective method to select the overall design scheme of initial product in parallel project, but the method must carry out a series of complicated derivation process. However, the neural network learning method proposed here can obtain accurate membership values ​​of other schemes of similar products as long as the fuzzy decision results of the existing schemes are studied. The early multi-objective decision making was developed through YP-CIMS, confirming the accuracy and effectiveness of this method.