Geographic Information System (GIS) has been widely applied to the field of pest management, which provides the new ways and methods for integrated pest research, prevention and control. The main purpose of this study is to forecast and manage forest pests and diseases. For this purpose, I developed the forest pests diseases forecast system, by using ArcObjects (AO) and Visual Basic as the secondary development platform. The system is able to synthetically analyze and judge relevant data, such as weather forecast and tree species, according to the occurrence laws of the forest pests and diseases and the characteristics of biology. The system uses Markov chain and other methods to forecast the occurrence period, amount, scope and the degree of harm of pests and diseases. To validate the reliability of Markov chain model, the pests and diseases data of Liu'an City of Anhui Province, in China from 1975 to 2001, to be applied. The results show that the method used is reliable. The occurrence area and location of forest pests and diseases can be labelled and displayed on the electronic map with three-dimension.