Abstract: An M_S 5.5 earthquake struck Taxkorgan county, Xinjiang, China on May 11, 2017. The earthquake caused 8 death, 31injured, and all kinds of property losses, is the typical event so-called "disasters by small earthquake". The rapid assessed earthquake losses according to the general method, are quite different from the actual ones. Therefore, it is of great significance to carry out the comparative analysis of the assessments of the earthquake losses and to understand the reasons of the deviation of the evaluation results in order to improve the assessment method. This paper analyzes the effect of earthquake disaster (seismic intensity distribution) and the population distribution etc. on the evaluation of earthquake life loss. The results show that seismic intensity distribution estimated according to elliptical seismic intensity attenuation model centered with macro-epicenter is of better estimation to the practical seismic intensity distribution than that with micro-epicenter. Another reason is that the area of seismic intensity region is significantly smaller than that of the practical one. We believe that because of the relatively high population density and relatively low anti-seismic capability of adobe and brick-wood structure buildings in the extreme disaster area, the unexpected high life loss was caused by such a earthquake of small magnitude. The comparative analysis results demonstrate that it will be helpful to improve the accuracy of emergency earthquake loss assessment through improvement of the high quality data of population, building and other risk related, as well as the accuracy of fast estimated earthquake location and seismic intensity distribution.