The Loess Plateau is well known for its severe soil erosion. Soil erosion severely interferes with the process of plant development and succession. However, plants do survive in this region of very serious soil erosion. These individuals can overcome stress and disturbance due to soil erosion through various breeding strategies, as well as through morphological and physiological compensation. Therefore, a plant is considered erosion-resistant when it not only has survived and adapted to the soil erosion environment, but it can also protect and improve the soil, prevent further soil erosion, regenerate itself, and maintain the plant community's stability and sustainable development. This study was based on several years (2003-2014) of vegetation survey data of the Yan river basin in the hill-gully region of the Loess Plateau. The research objective was to select potential erosion-resistant species, and subsequently illustrate their distribution characteristics and their survival and propagation ability. The definition of erosion-resistant species and Braun-Blanquet phytosociology methods were used to select potential erosion-resistant species. The main results were as follows:1) A total of 42 potential erosion-resistant species were selected, which belong to 18 families and 33 genera. Gramineae, Leguminosae, Compositae, and Rosaceae species accounted for 66% of the total potential erosion-resistant species. 2) Phanerophyte, chamaephytes, and hemicryptophytes were the main plant life forms, and they accounted for 85% of the total erosion-resistant species. Shrubs or small shrubs, and perennial herbs were the main plant growth forms, and they accounted for 76% of the total erosion-resistant species. Xerophyte and mesophyte plants were the main plant water ecological types, accounting for 78% of the total erosion-resistant species. Furthermore, combined with the climatic conditions of study area, the 42 potential erosion-resistant species could be divided into three types:eurytopic species, medium amplitude species, and stenotic species. 3) Overall, 55% of the species were erosion-resistant, having a maximum cover over 50%, and could be the structural or single dominant species in a community. Maximum cover of the other species was less than 50%, and these species usually were the co-dominant species. All potential erosion-resistant species had a high cover and aboveground biomass, indicating that these species could adapt to different erosion environments and grow adequately. 4) Almost all 42 potential erosion-resistant species had a soil seed bank and seedling emergence, and 60% erosion-resistant species had canopy seed bank. Except for annual plants, allspecies were able to reproduce asexually. This result indicated that the potential erosion-resistant species could maintain their own survival and reproduction. 5) From the 42 potential erosion-resistant species, 13 had a basal stem-shoot architecture, with a large crown that can protect the base soil. Furthermore, eight species have an expanding shoot architecture, which display a strong ability to protect the soil and intercept sediments. In addition, six species have dense shoot architecture, and seven have a tussock-forming shoot architecture. Both types can effectively intercept sediments. In conclusion, the potential erosion-resistant species, which only occupied 13% of the total species recorded, had a seed and seedling bank. Furthermore, the perennial plants mainly employed vegetative reproduction, and could maintain their own survival and reproduction. Because the plant crowns can protect the soil surface and plant base stems can intercept sediments, a mound can be formed in the plant base, which will effectively control soil erosion.