In Morocco we tested the consistency between an a priori habitat classification based on topography, hydrology, vegetation structure and composition, and an a posteriori classification based on arthropod assemblages, in a plain supporting wild endangered Houbara Bustards. According to vegetation structure, we defined seven a priori habitats that differed significantly in perennial cover and height. A multivariate multiple regression analysis showed a significant relationship between arthropod assemblages and vegetation structure. Canonical Analyses of Principal Coordinates, conducted simultaneously on direct searches of arthropods and trapping data, showed significant differences between assemblages in both cases, and produced two similar constrained ordinations of six a posteriori habitats: esparto grass (Stipa tenacissima), temporarily flooded areas, fields, “reg” with short perennials, “reg” with tall perennials and wadis. The two sampling methods reflected a dominance of ants and beetles. Arthropod biomasses increased significantly in spring and then decreased significantly in summer for beetles, and in autumn for ants. No strong differences appeared between habitats within seasons, especially in spring, indicating a uniform distribution of food resources during the Houbara breeding season. The “reg” with short perennials had the highest ant biomass in summer. This “reg” and fields also supported the highest arthropod biomass in autumn. Variation in arthropod biomass was a pertinent factor that should be integrated into Houbara habitat selection studies. The definition of habitat availability based on easily identifiable landscape units, combined with empirical tests on arthropod communities provided an accurate classification for habitat selection studies and conservation planning.