Abstract: The effective prediction of the landslide deformation plays a key role in the prevention and formulation safety measures of engineering projects. The landslide hazards are frequently encountered, and the triggering mechanism is extremely complex, therefore, the time of landslide occurrence is difficult to predict. In the view of the shortcomings of the current single deformation prediction model, this paper proposes a combined and weight gray-trend optimal model GM (2,1) for the landslide prediction based on the accumulate deformation data of Shuping Landslide from June 2004 to June 2013. The results show that the combination forecasting method has the highest prediction accuracy and adaptability and is closest to the actual monitoring results. Therefore, the combined prediction model is stable and reasonable, it can be used to predict the landslide deformation in the reservoir area and even other areas.