道路交通事故预测是交通研究的一个重要课题.以我国交通安全状况为研究对象,依据我国道路交通事故的特点,利用神经网络具有自学习、自组织、自适应能力特性,运用神经网络的方法及我国多个年度道路交通事故统计数据,建立了道路交通事故神经网络宏观预测模型,预测精度符合道路交通事故预测的要求. Road traffic accident prediction is an important issue in traffic research.According to the traffic safety in our country, based on the characteristics of road traffic accidents in our country, this paper makes use of neural network’s self-learning, self-organizing and self-adaptive ability features, The statistical data of many annual road traffic accidents in our country set up the macro prediction model of road traffic accident neural network, and the prediction accuracy accords with the requirement of road traffic accident prediction.