由于小波神经网络对建筑物变形预报具有较高的模型拟合及预报精度,文章从小波神经网络算法原理出发,阐述了使用此方法对所获得的基坑变形监测数据进行模型建立及预报的过程,并运用C++语言实现了编程,经过对某基坑变形监测预报的研究表明,该方法具有很强的可行性和实用性,可以及早为基坑梁变形做出预警,以避免或减少灾害的发生。 Because wavelet neural network has a good model fitting and forecast accuracy for building deformation forecasting, this paper describes the process of establishing and forecasting the obtained deformation monitoring data of foundation pit by using the principle of wavelet neural network algorithm , And the programming is implemented in C ++ language. After the monitoring and forecasting of a pit deformation, it is proved that this method is very feasible and practicable, so as to early warning the deformation of the foundation beam in order to avoid or reduce the disaster occur.