Abstract Because in deposit 2 of Jinchuan Company,the deep stope area will reach 10×104 m2 and the buried depth will reach 1 000 m,the key point of the second stage project is the stability of deep wall rock and filling. At the same time,the deformation feature of wall rock and filling also affects the project’s safety and normal operation. A great deal of deformation data are obtained in Jinchuan mine by many methods,such as beach mark,convergence and GPS. Based on the deformation data,the deformation rule of wall rock and filling is predicted with neural network model,and some useful conclusion is obtained.