以实测数据为基础，在精轧预设定中采用ＢＰ神经网络的方法取代传统的轧制力数学模型，并对神经网络输入项和训练样本进行分析，提出改善神经网络预报精度的一些方法。预报结果和实测数据比较表明，预报精度有较大的提高。为满足现场要求提出神经网络预报轧制力在线应用方案。 Based on the measured data, the traditional rolling force mathematical model was replaced by the BP neural network in the presetting of finishing rolling, and the input items and training samples of the neural network were analyzed. Some methods to improve the prediction accuracy of the neural network were put forward. The comparison between the forecast result and the measured data shows that the prediction accuracy is greatly improved. In order to meet the requirements of the site, neural network prediction rolling force on-line application program.