In arid and semi-arid regions, groundwater availability is one of the controls on vegetation distribution. This groundwater-dependent distribution of vegetation has been particularly observed in the Hailiutu River catchment, a semi-arid region in North China. We used remote sensing images of vegetation index (normalized difference vegetation index, NDVI) and field data of depth to water table (DWT) to assess the response of vegetation distribution on increase of DWT at the regional scale. The frequency distribution curves of NDVI with respect to different DWT were obtained. The statistical distributions of NDVI values at different DWT intervals indicate that higher vegetation coverage and more plant diversity exist at places of shallow groundwater. Both the mean and the standard deviation of NDVI values decrease with the increase of groundwater depth when DWT is less than 10 m. Beyond that depth, a low level of vegetation coverage and diversity is maintained. Comparisons of different sub-areas within the region with different dominant species showed that the NDVI of shrubs is sensitive to DWT. In contrast, NDVI of herbs is not significantly influenced by DWT. The relationship between NDVI and groundwater depth in farmlands could not be reliably determined because of disturbance by human activities. We conclude that application of this methodology may significantly improve our ability on sustainable management of land and groundwater resources. Copyright © 2012 John Wiley & Sons, Ltd.