Earthquake is one of the most serious natural disasters that has caused great casualties in China. The rapid assessment of casualties after earthquake is the key to earthquake emergency response. The pre-assessment survey before earthquake is an effective means to improve the accuracy of post-earthquake rapid assessment. Through analysis of 62 historical earthquakes in the western region of China,we found that the earthquakes with magnitude less than 4.5 would not cause death. In this study we used the pre-evaluation data of some districts and counties in Yunnan and Sichuan Province to develop an exponential function relationship between the number of deaths and the magnitude of the earthquake,to construct the exponential function relationship for each district and county,and then to calculate the regression coefficient. Based on the above results,the relations of earthquake casualties with magnitude 5.0 to 8.0 have been obtained at intervals of magnitude 0.5,which can be used to assist the rapid assessment after the earthquake.