Abstract: Based on GIS techniques and a climatology model using several types of geographic data such as land-use information, longitude, latitude, altitude, slope grade and direction, etc., cold-disaster data are corrected according to specific geographic details. Considering growing conditions and cold-index of winter economic fruit trees, a method is presented, by which when the cold-disaster occurs, its developing condition, intensity and area can be watched and predicted in real time and dynamically. The key techniques and conclusions are: (1) The cold disaster index is defined based on two factors (temperature dropping range and extreme lower temperature during the course) to offset the disadvantage of the conventional operation system. (2) Based on the output of Numerical Forecast Model T213 (developed in China) and the MOS (Model Output Statistics) forecast equation, cold disaster forecast for 86 ground meteorological stations in Guangdong Province is made. (3) The geographical correction model considering slope grade and slope and slope direction information together with three general factors (longitude, latitude and altitude) is adopted and get a good result.