Qinling zokors (Myospalax rufescens) had been studied at Pingheliang Natural Reserve in Shanxi Province during June to September in 2009. According to the local situations, 13 habitat factors were considered in fixed line-sampling. 302 samples have been set up with a spacing of 5 m, and 93 new cave-sites being used by zokors were also found and sampled. It is shown that vegetation-type selectivity is determined by zokor's feeding preference. Qinling zokors select grass and radish sites, avoid the wheat, bamboo and shrub sites, and randomly use potato and maize sites. Although zokor's cave-paths are segregated with deep soil layers, which are hardly affected by wind, frost, rain and dew, zokors still select the appropriate cave-hatches in order to push soil outside easily or perform their outside-cave activities expediently. Zokors mainly select sunny slopes as their cave-hatches, avoid half-gloomy and half-sunny slope and randomly select gloomy slope. Sunny slopes have short water-collecting time if it rains, and receive more sunshine than other slopes, so they are beneficial to zokor's activities. Especially, flourishing plants show such high diversity that their roots provide rich nutrients in sunny slopes, which are favorable to zokor's survival and multiplication. It is found that soil rigidity and stone-percentage determine the direction of zokor cave-path and selection of cave-sites. In terms of effects of soil factors, Qinling zokors select their cave-sites in soft and low stone-percentage soils, and avoid hard and high stone-percentage soils. Human disturbance is often an important factor affecting the habitat selection of many animals. The cave-site selection of Qinling zokors is relatively different from other animals in that they prefer intermediate human-disturbed sites, avoid sites without human disturbance, and randomly selected strongly disturbed sites. Principle component analyses indicated that the most important factors influencing zokor's cave-site selection include (in order of decreasing importance): distance to water resource, distance to residence, distance to highway, herbage density, herbage cover, shrub cover and shrub density. In conclusion, zokors are inclined to select cave-sites with sparse shrubs and close to highway, but other habitat factors (i.e., distance to water resource, distance to residence, herbage density and herbage cover) have nearly no influence on zokor's cave-site selection. Based on the aforementioned results, two suggestions are proposed for prevention and control of zokors. Firstly, intermediate-cutting strategies should be used in the process of forest management, and young seedlings should be planted as densely as possible; for reforestation of grass sites, some bulky trees (especially poisonous trees to zokors) should be planted at first, then young seedlings need to be reinforced, and stone-percentage of soil should be improved through laying some big stones in the soil after the trees survive. Secondly, multi-cropping strategies should be used in the process of agricultural production, and some crops with developed root systems (or poisonous crops to zokors) should be intercropped with the main crops. Moreover, weeds should be removed frequently, which not only benefits the growth of crops, but also is an effective means for rodent pest control.