The concave-eared frog, Odorrana tormota, is a rare species native to eastern China and the first non-mammalian vertebrate demonstrated to both produce and perceive ultrasonic frequencies. In the current study, for the first time, we reported the geographic distribution, habitat selection, and relationship between different disturbance intensity and relative population density of the species in southern Anhui Province. A total of 32 different mountain streams were investigated from April 2012 to August 2013. Among these investigated streams, the species was detected in 16 streams. Except for Xiangxi and Fuxi in Huangshan District, this is the first record of the species in the remaining14 streams. The distribution range of the species within the stream varied among different streams from the shortest range less than 1 km to the longest range of more than 20 km. The 16 mountain streams are the tributaries of Qinyi, Xin'an, or Shuiyang river, and the species was not found in any tributary of other rivers (such as Qiupu and Qimen river, etc.) in the region. These results suggested that at present, the concave-eared frog of southern Anhui Province have survived in some isolated mountain streams linked by Qinyi, Xin'an, or Shuiyang rivers. In order to reveal the strategies of habitat selection of the species, we tested the differences in frog numbers among six different habitat types including trees, shrubs, or herbaceous plants along river banks, benchland and rocks or other objects above the water. In total, 222 individuals from 10 different mountain streams were analyzed and significant difference in number of frogs was detected (F_5,54=25.75, P < 0.001) in herbaceous vegetation and shrubland, which harbored significantly more individuals than other habitat type, suggesting that the two habitat types along mountain streams play an important role in species' proper survival and reproduction during the nighttime. It should be noted that, although the recorded frog number in arbors was significantly less than in herbaceous vegetation and shrubland during the surveyed period, arbors might provide indispensable hideout place for the species during the daytime. Additionally, we evaluated the impact of human disturbance on relative population density of the species. Initially, we obtained the relative population density for each of the 16 streams through a calling-based method. The number of male frogs' calls in the 100-m range along each stream was recorded and used as the value of relative population density. We classified human disturbance on a three-point scale: strong, medium, and weak according to the degree of destruction as a result of human activity on the habitat of the species. One-way analysis of variance (ANOVA) was used to test the effects of human disturbance on relative population density. The results showed that human disturbances significantly affected population density (F_2,13=8.155, P=0.005). The least significant difference (LSD) tests revealed that relative population density under medium disturbance intensity was the highest. Our findings suggested that moderate anthropogenic interference might increase the species number and quantity of insects and then attract more frogs. However, with the increase of disturbance intensity, the damage to vegetation along the stream is more intense causing dramatic decrease in population density.