The historic observation data of dry-hot wind of spring wheat in Yongning, Ningxia and the wheat yield recorded by Ningxia Statistics Bureau are analyzed with the aid of disaster data got from Civil Administration Department of Ningxia. The clustering analysis method sieving step by step and giving an original clustering kernel as well as a segregated method step by step of the spring wheat yields are used. The indexes for two types of dry-hot wind of spring wheat in the Ningxia irrigated areas are obtained. Then, the discriminate equations are established respectively in the different disaster grades. The yield losses of spring wheat suffered from dry-hot wind in the grouting stage are segregated by means of climate simulation before the flowering and tasselling stage. A synthetically checking table for different disaster grades is given to combining the influence between the yield loss and the disaster degrees suffered from dry-hot wind in the different wheat growth stages. An estimating yield loss model is established and the monitoring disaster and the estimating yield losses are available .