A web-based prediction system for wheat stripe rust was developed based on B/S (Browser/Server) mode in this study. Some existing prediction models of wheat stripe rust were collected, analyzed and then stored in SQL Server 2005 database according to certain rules. All these models could be called through this web-based system and used to predict wheat stripe rust. Meanwhile, Using multiple regression analysis principle, prediction regression model could be built based on the input historical data of wheat stripe rust through the network programming via this system, and significance tests of prediction factors could be conducted to obtain optimal prediction model and the built model could be stored into the model database for further prediction of this disease. Using WebGIS technologies, the prediction results of wheat stripe rust could be displayed in different colors in the web map according to the prediction values of disease prevalence. The web-based prediction system for wheat stripe rust developed in this study provided a convenient and fast way for the prediction of wheat stripe rust.