Abstract Natural factors such as water resources, climate, topography, etc… play an important role in site selection specifically rural settlement where people's way of living mostly depends on the environment. In this study we classified one of the sub-catchments of the central desert of Iran, called Bardaskan and analyzed the local position of villages. During this process, first the effective parameters in site selection of villages were determined and gathered in ArcGIS 10 and the relevant maps were developed. Then, defining fuzzy operator, the fuzzy inference network is developed in order to obtain the final image. By evaluating the relevant criteria, the region was divided into 5 classes of relatively weak, weak, intermediate, relatively proper, and proper. Finally, the current position of settlements in proportion to these classes is assessed. The results of findings showed that from a total of 71 villages, 36 were in the relatively weak class, which means 50.7 percent. But 54.92 percent were in the intermediate class. As the area increases, more unfavorable environmental situations are observed. 2000 square kilometers (80 percent) of the area out of the total area (2500 square kilometers) are in the weak to relatively weak classes. Keywords: fuzzy logic, Rural Settlements, village, central desert, Bardaskan