为了提高变形抗力预测精度 ,针对传统热轧变形抗力模型的固有缺陷 ,提出一种将动态 BP神经网络与数学模型相结合的新方法。大量实验数据表明 ,采用这种方法 ,预测结果的相对误差在± 6%以内 ,预测精度优于传统方法 In order to improve the prediction accuracy of deformation resistance, a new method of combining dynamic BP neural network with mathematical model is proposed in view of the inherent defects of the conventional hot deformation resistance model. A large number of experimental data show that using this method, the relative error of the prediction results within ± 6%, the prediction accuracy is better than the traditional method