Resumo. The classification techniques, such as cluster analysis, used in biological areas, in the plants and animals classification keys, are also useful for defining homogeneous regions, for climate data or for identification or monitoring of vegetation. The objective of the research was to characterize the spatiotemporal dynamic of NDVI (Normalized Vegetation Index Difference) of vegetation groups of Rio Grande do Sul, clustering homogeneous regions based on NDVI temporal variability and to analyze the pattern of NDVI annual variability of the different groups. For this NDVI data from the Global Inventory Modeling and Mapping Studies (GIMMS) series, derived from Advanced Very High Resolution Radiometer / National Oceanic and Atmospheric Administration (AVHRR/NOAA) over the period from January 1982 to December 2008 were used. Six groups were defined, which were compared to environmental macrozoning. The results showed that the variability of NDVI, obtained from GIMMS/AVHRR series, allows clustering homogeneous regions in order to identify and map the main vegetation groups that occur in the Rio Grande do Sul State. Vegetation groups are characterized by a typical seasonal pattern, however, with amplitude variations of NDVI between them, which allow to monitor phenological changes in vegetation due to its temporal dynamics.