Abstract: In order to improve the prediction accuracy of the GM（1,1）model for its application in residual deformation prediction of goaf, the residual modification GM（1，1） model was constructed based on the gray GM（1，1） model. The positive residual fitting correction of GM（1，1） model prediction results was carried out with unary quadratic polynomial. The monitoring data sets of surface residual settlement in goaf with equal and non-equal spacing were used for example verification. The results show that: the prediction accuracy of GM（1，1） model modified by nonlinear polynomial fitting after residual correction is improved obviously. However, GM（1，1） model modified by polynomial fitting of original residual error has no obvious improvement effect on prediction accuracy. Residual positive fitting correction is an effective post-processing improvement method for GＭ（1，1） model.