Abstract: Based on the survey data of 39 samples of winter wheat stripe disease field rate in 8 counties of 5 cities in Hedong of Gansu, the meteorological grade models of incidence area in autumn and spring of wheat stripe rust were established by statistical method. This paper analyzed the development causes of wheat stripe rust in Hedong of Gansu and provided accurate forecasting techniques, so as to provide scientific basis for timely disaster prevention and mitigation in agriculture production and reducing economic losses. The results showed that: the wheat rust over summer was at the 8-day moving average temperature below 24-25℃ in the hottest days of summer, or wheat rust infection drifted from other places with wind, and with the decrease of temperature in autumn and the increase of atmospheric humidity, the suitable temperature and humidity climate caused the wheat rust quickly multiplying and spreading, resulting in the rapid development of local wheat stripe rust; after entering the winter, the rust overwintering was at the 7- day moving average temperature above - 9- -10℃in the coldest days of winter, and the rising temperature and the increasing precipitation in spring caused large area spread of wheat rust and rapid infection to the downstream area. Statistical meteorological grade forecasting model was tested by the back substitution, the accuracy in autumn and spring was 89% and 91%, respectively, the forecasting accuracy was relatively high.