Abstract: In order to get an accurate estimate of the landslide displacement nonlinear changing trends under the small-data situation, a model based on the pGM (1,1) and generalized regression neural network was proposed. The new model integrates the advantages of the generalized regression neural network learning speed, high prediction accuracy and the pGM (1,1) model of reducing of the data randomness and enhancing the regularity. Comparing two engineering examples using the new model with the results of the previous studies, it was shown that the new method was feasible and effective.