Abstract: In order to construct the hierarchical prevention system of termite hazard in dams of Yunnan Province and realize the accurate prevention and control of termite hazards, the statistical method of termite hazard numbers was optimized by using the class Ⅲ disease normalization algorithm, and an evaluation method of the original termite hazard density of dam was proposed. Six factors (mean annual temperature, relative humidity, total water resources, altitude, soil conditions and food condition) affecting termite hazard were selected, and the entropy weight-TOPSIS method was used to determine the weight of each influencing factor and their comprehensive score on termite hazard. The results showed that the termite damage of reservoirs in Yunnan Province was more serious in the south and southwest, and relative mild in the northwest and northeast. The comprehensive score was highly consistent with the original termite damage density, which verified the reliability of the score results. Based on this, a three-level prevention system for general, medium and key prevention areas was established according to different score intervals, and corresponding prevention measures were formulated. The research results can provide scientific guidance for the accurate prevention and control of termite damage in reservoirs of Yunnan Province, and also have certain reference significance for termite prevention in other regions.