In ecosystems, habitat patches are rarely isolated but embedded within surrounding landscape matrix. The resources and environmental conditions required by the species population not only depended on the habitats, but also on the composition and structure of landscape matrix surrounding the habitats. Black-necked crane is a flagship species on the Qinghai Tibet Plateau, and the alpine wetland along the valley of the middle reaches of the Yarlung Zangbo River is the world largest overwintering habitat providing a high-quality foraging environments for the species. Clarifying how the selection of foraging habitats of black-necked crane is influenced by landscape matrix is of great significance for flagship species protection and comprehensive management of watershed ecosystems. Based on landscape ecology principles, this study used the data from remote sensing images and field survey. By constructing multiple-scale buffer zones of landscape matrix, conducting multivariable correlation analyses, and building the Maxent model, spatial and temporal changes of landscape patterns and habitat characteristics, and their relatedness with black-necked crane population along the valley of the middle reaches of the Yarlung Zangbo River from 2000 to 2020 were analyzed. And then, the influences of landscape matrix structure on the selection of foraging sites of black-necked cranes were investigated. Further, the key habitat factors were used to simulate the habitat suitability distribution of the species. We found that:(1) Temporally, the area of croplands along the river valley increased during early years but decreased during recent years, and shallows continually decreased; Spatially, species population demonstrated a cluster distribution pattern concentrating on the east but less on the west, which overlaps with the distribution of cultivated land along the valley; (2) Landscape matrix significantly influenced the selection of foraging habitats. Structurally, black-necked cranes prefer foraging habitats surrounded by the matrix with high landscape connectivity and high dominance in landscape structure; compositionally, the matrix with cultivated land types are mostly preferred. If the matrix with high proportion of other landscape types or fragment cultivated-land patches, the species would not select the habitats within such matrix, which may relate to the foraging habits and the demand for environmental security of black-necked cranes; (3) The influences of structure and composition of landscape matrix on habitat selection of the species also showed scale effects. The most significantly influencing scale is 1500-2000 m, but different landscape types, such as cultivated land, grassland, and water area, showed different influence scales at 1500m, 3000 m and 4000 m, respectively; (4) Through modeling simulation, it is revealed that the suitable habitat area of black-necked crane increases at first and then decreases, but it shows an upward trend as a whole compared with that in 2000. The distance of the habitats to cropland, patch density of waters and landscape dominance were always the most contributing landscape factors in explaining the habitat suitability. The study reveals that, the rural planning in this region should optimize black necked crane's habitats and surrounding landscape matrix in which crop production and household living would be considered as a whole, forming a comprehensive management of the alpine wetland ecosystems taking the protection of the flagship species as the core. Such a management would enhance the stability and sustainability of the alpine ecosystems on the Qinghai Tibet Plateau. Meanwhile, results of this study also provide a perspective for in-depth studies on the mechanism of species habitat selection.