In recent years, At home and abroad scholars proposed a great deal of mathematical models on landslide forecast, which through the fitting of landslide monitoring data and trend analysis to determine the landslides time. How to determine the results of the analysis to different mathematical models and identify good and bad quality of the forecast model is an important problem to the landslide forecast and decision-makers. The author proposed fitting effect index(including posterior index, model fitting efficiency index and RMS error) and experiment forecast effect index(including experiment forecast index and related coefficient index), which establishment quality testing model of landslide forecast. Using gray GM (1,1) mathematical model, the three exponential smoothing model and time series model fit and experiment forecast the monitoring data of Lianziya dangerous rock body in the three gorges of the Yangtze river. Using fitting effect index and experiment forecast effect index comprehensive analysis, the results showed the quality testing model of landslide forecast is an effective and practical approach. Landslide forecast research has become a hot topic from the 1960s (Xu et al., 2004; Yang et al., 2004;. Li et al., 1999; Huang et al., 2004; Yin. 2004; Wen et al., 2004). In recent years, Experts and scholars from home and abroad have done a lot of research work at the time of landslide forecasting, which made a variety of landslide forecasting model, such as the gray forecast model(Li et al., 2007; Lu et al., 2001; Wang et al., 2005), neural network model (Lin et al., 2002; K.M. et al., 2004), and so on(Li et al., 1996; Yi et al., 2007; Xie et al., 2005). Various forecasting methods has its own advantages, but also there is a corresponding shortcomings and deficiencies. Most forecasting models are based on the monitoring of landslide displacement data were fitted speculate as to how to assess the quality of the forecasting model is good or bad, it becomes a problem for landslide forecast researchers and policy-makers to be solved. A good forecasting model is not only to be able to describe the past events, more important it is to be able to predict future events. Thus, for quality assessment of landslide forecasting model it is absolutely necessary. We can assess the quality of the model, in order to truly establish forecasting model in line with the actual conditions of the landslide, and reliable forecasting results. In this paper, by using the fitting and forecasting performance metrics to fully assess the quality of both the effects of landslide prediction model.
