Abstract: Deformation monitoring is one of the main contents in slope monitoring. Deformation prediction is one of the main technical problems. The current deformation prediction models have limitations. For example, the neural network predicting model needs massive data as study pieces and grey model requests that the original data must meet the exponent rule. A back propagation neural network prediction model and grey system GM(1,1) model for slope deformation are built. In order to gain more accurate prediction results, a weighted function combining forecasting model of neural network and grey system on slope deformation is developed and a dynamic programming method is used to solve multi-step decision problem, which can minimize the square sum of the relative errors and gain the combination weight. So, the predicting result can greatly enhance the precision and make up the limitation of single method. Through calculating an actual slope the predicting results develop greatly and consistent with the monitoring data. The combining predicting model enriches the slope deformation prediction theory.