Abstract: Under the background of frequent global warming and extreme climate events,people pay more attention to the disaster risk of agricultural and forestry production.It is of great significance to construct disaster early warning indicators,and carry out risk early warning and advance prevention for effective disaster prevention,mitigation and loss reduction.Based on the data of frost disaster situation,meteorological factors and planting area of fragrant pear from 1981 to 2020 in Xinjiang,the frost disaster risk index,exposure and vulnerability index were calculated,the frost disaster risk assessment model of flower period was comprehensively constructed,and the fragrant pear frost risk warning was conducted based on the grid forecast,verifying the applicability of the assessment model.The results show that the lowest temperature,process temperature range and low temperature duration are the main factors of frost disaster.The spatial distribution of frost risk in spring is generally high in the north and low in the south.The high risk areas of spring frost are mainly distributed in the western part of northern Xinjiang,the western and central regions of the northern slope of Tianshan Mountains,the northern part of Bazhou,Aksu city and its southern parts,with high risk and loss of frost.Based on the meteorological situation and the orchard disaster survey results,it shows that the combination of the frost risk assessment model and the lattice forecast can better forecast the risk zoning and influence level of fragrant pear frost,which is basically consistent with the actual occurrence area and percentage of the frost disaster of fragrant pear.The frost index and risk warning model are reasonable and applicable.