Eremosparton songoricum is a rare desert perennial shrub found only in Central Asia. In China, its distribution is limited to the Gurbantunggut Desert. Eremosparton songoricum is a pioneer species that can become established on bare sand, where it acts as a sand stabilizer, and also occurs in artificial sand fixed areas, created using sand fences. In the sand fixed areas, the increased stability from the fences leads to the existence of biological soil crusts. This further stabilizes the surface physical structure and enhances the soil nutrient content, thus creating favorable plant growth conditions which, in turn, increase plant species diversity. In contrast, the natural bare sand is barren with high mobility due to high winds and thus no stability develops to create plant growth conditions. The spatial distribution patterns of E. songoricum were investigated in both bare sand areas and artificial sand fixed areas by plotting coordinates of every E. songoricum ramet in each sample plot. To assess the quantitative population characteristic differences of E. songoricum between artificial sand fixed areas and natural bare sand areas, species composition, abundance, height, coverage and biomass were assessed in both habitats. Results showed that the numbers of families, genera and species of E. songoricum communities in the artificial sand fixed habitat were significantly higher than those in the bare sand habitat. The similarity index value between communities in the two habitats was only 0.522, indicating that the artificial sand fixed habitat changed the species composition and structure of the E. songoricum community greatly. The quantitative characteristics of height, coverage, above ground, below ground and total biomass densities and above ground, below ground and total biomass per individual were also significantly higher (P <0.05 or P <0.01) in the artificial sand fixed habitat than those in the bare sand habitat. The correlation dimension values of the population spatial distribution in the two habitats were all close to 2.0, indicating strong spatial relevance and holding space ability for individuals. The aggregation analysis at two scales (1 m and 2 m) indicated that almost all of the populations were randomly distributed; thus the artificial sand fixed habitat had no effect on the degree of spatial aggregation. Random distribution pattern may therefore be part of the intrinsic properties of E. songoricum populations and it is possible that it forms the same way regardless of habitat. The organic matter, total nitrogen and available nitrogen contents in the artificial sand fixed habitat were significantly (P <0.01) higher than those in the bare sand habitat, and were significantly (P <0.05 or P <0.01) positively correlated with the population quantitative characteristics except for abundance of E. songoricum individuals. In comparison with the bare sand habitat, the surface of the artificial sand fixed habitat was stable and the soil organic matter and nitrogen contents were higher. This meant that the coverage, height, biomass densities and biomass per individual of the E. songoricum population were markedly different and that the survival and adaptation ability of E. songoricum is stronger in these plots. This conclusion is important for the protection and conservation of this and other endangered desert species.