Xiaolongshan Forest is located in the transition region of temperate zone to subtropical zone which is also the junction area of four natural vegetation areas in China. There are lots of kinds of floristics and big differences in landform and climate in the forest. Forestry biological disasters happen frequently and the environmental climatic conditions are main limiting factors for biological disasters. The technique of forecasting biological disasters to protect the forest has been studied, which is of significance for forest development. The method of stepwise regression was used and the model for forecasting the zekor damage area in spring and autumn and statistical model were established. Zekor likes to live under warm and dry condition and dislikes light, rain, high temperature and high moisture. Zekor reproduces well in warm winter, in warm spring and autumn with little rain, and in sunshine and suitable temperature in early summer. The result of the model for forecasting showed that average prediction accuracy in spring and autumn was above 97% in the past 16 years. We tried to forecast the zekor harmful area in 2008, 2009 and 2010 with the model. The result showed that the average error rate was 1.8,2.0 and 1.9,1.6 in 2008, 2009 and 2010 respectively and the prediction accuracy was above 98%. The effect of forecast reached ideal level. We supposed that the models would meet the need of professional and service work.