Normalized difference vegetation index (NDVI) derived from SPOT HRV multispectral data was used to study the changing environmental quality of Hong Kong from 1987, 1991 and 1993 to 1995. Conventional change detection techniques such as image differencing or principal components analysis helped to highlight salient changes. These techniques, however, were less effective in identifying subtle changes, in particular the amount and quality of green space. Integrating the mean NDVI values at the Tertiary Planning Unit (TPU) level with census and land-cover data showed that the NDVI values were related to woodland, tall scrubland and high-density urban areas. It was also related to the level of crowding as depicted from a factor analysis of census data. Tracing the changing pattern of mean NDVI values revealed that areas with continuous increases in NDVI values are scattered around old urban districts experiencing improved landscaping. Areas of continuous decrease in NDVI values covered a large part of rural New Territories and western Hong Kong Island revealing the urban expansion process. This provided valuable information for the assessment of environmental quality for planning and management of the environment.