Abstract: In order to evaluate grassland loss caused by Calliptamus abbreviates, a study has been undertaken to establish the correlation between the Normalized Difference Vegetation Index (NDVI) of Leymus chinensis grassland, as measured with a high resolution radiometer, and the density of C. abbreviates populations. Five groups of C. abbreviates with densities of 5, 10, 20, 40 and 60 heads/m² were used to simulate the damage caused by locusts feeding in the field. NDVI was measured after different damage periods and losses caused by the locusts were calculated using a regression model of NDVI and grassland biomass. Results indicated that NDVI decreased with increases in locust density. However, at a density of 10 heads/m² the NDVI rose slightly. The regression equations of NDVI and locust density under different damage periods were: y=0.5932+0.0014x-6.93×10^-5x² (5 days), y=0.5950-4.8500×10^-4x-4.01×10^-5x² (10 days), y=0.5848-0.0024x-1.61×10^-5x² (15 days) and y=0.6422-0.0031x-2.12×10^-5x² (20 days); where y represents the NDVI score and x locust density. The study found that in the case of low densities (less than 20 heads/m²) the damage caused by C. abbreviates had no obvious effect on NDVI with the extension of damage time. However, under high densities (more than 20 heads/m²) the NDVI decreased rapidly. The regression model (y=614.15x-119.28) shows that NDVI data can be used to measure grassland loss. The study indicated that low densities of C. abbreviates (5 and10 heads/m²) had significantly less effect on L. chinensis grassland but that when locust density reached 40 heads/m² the biomass decreased sharply. Results thus showed that the hyper spectral grassland vegetation index (NDVI) and the damage caused by locust density are significantly correlated. The spectral change caused by locust damage can thus be used to estimate both the density of locusts and the extent of grassland loss. This research provides a foundation of further large-scale grassland monitoring with hyper spectral remote sensing.