The wheat scab is one of the most serious disasters affecting wheat yield. It is important to forecast the yield loss of wheat scab for the agricultural production and the national foodstuff security. The method was developed dynamically to forecast yield loss of wheat scab by using the data of yield loss of wheat scab and ten-days average temperature, ten-days precipitation, ten-days sunshine, ten-days precipitation days and the index of composite cluster analysis. The results show that the accuracy of tendency forecast is 90%, and the average accuracy of quantitative forecast is 755% for 10 years. This method is simple operation with high accuracy and good prospects.