化学品船运事故分析是船舶应急事故处理的一项重要内容。广义回归神经网络(General Regression Neutral Network,GRNN)是建立在非线性回归原理上的一种径向基网络,在样本数据较少时也有较好的预测结果。在相关数据的基础上,分析了事故各因素及彼此间的相互影响,利用GRNN建立了相关模型,训练样本用来训练模型,其余样本用来进行预测。结果表明,GRNN具有较高的准确度及稳定性。 Chemical ship accident analysis is an important part of ship emergency management. General Regression Neutral Network (GRNN) is a radial basis network based on the principle of non-linear regression. It also has good prediction results when the sample data is small. Based on the relevant data, the factors of the accident and the interaction between them were analyzed. Relevant models were established by using GRNN. The training samples were used to train the models and the rest samples were used for the prediction. The results show that GRNN has high accuracy and stability.