With BP nerve network, the experimental data are studied and predicted to optimize the composition of short fibers mixed in the flexibility agent of oil-well cement, improve the veracity of selecting materials and provide the theoretical ground of selecting and applying fiber flexibility agents. 文章通过BP神经网络对实验数据进行学习预测,优化了混杂短纤维在油井水泥增韧剂中的组成,提高了选材的准确性,为纤维增韧剂的选择和应用提供了一定的理论依据。