The ground is intensively heterogeneous and fragmentized, and the vegetation is very sparse in arid region, so the selection of the vegetation index of vegetation degeneration measurement is the research hotspot on vegetation remote sensing. In this study, the project area of controlling the sources of dust storms affecting Beijing and Tianjin was selected as the research object. The vegetation indexes of vegetation degeneration measurement fitting the research area were studied from the tow aspects on reducing the impact of soil background and the ability of detecting low coverage vegetation. The results have shown that GEMI had a maximum effect to reduce the soil impact, and NDVI had the strongest ability to detect low coverage vegetation. This research deems that NDVI is the best selection to measure vegetation degeneration in arid region, and GEMI is the following selection.