ABSTRACT. Latin America in general and the Rio de la Plata Grasslands (RPG) in particular, are one of the regions in the world with the highest rates of change in land use/land cover (LULC) in recent times. Despite the magnitude of this change process, LULC descriptions in the RPG are far from being complete, even more those that evaluate LULC change through time. In this work we described LULC and its changes over time for the first 14 years of the 21st century and for the entire grassland biome of the Rio de la Plata, one of the most extensive grassland regions in the world. We performed simple but exhaustive classifications at regional level based on vegetation phenology, using extensive LULC field database, time series of MODIS NDVI satellite images and decision trees classifiers, generating an annual map for all RPG. The used technique achieved very good levels of accuracy at the regional (94.3%-95.5%) and sub-regional (78.2%-97.6%) scales, with commission and omission errors generally low (Min = 0.6, Max = 10.3, Median = 5.7, and Min = 0, Max = 41.8, Median = 6.8 for regional and sub regional classification respectively) and evenly distributed, but fails when LULC classifications are generated in years when the climate is very different from those used to generate spectral signatures and train decision trees, or when the NDVI time series accumulates large volumes of lost data. Our results show that the RPG are immersed in a strong process of land use change, mainly due to the advance of the agricultural frontier and at the expense of loss of grassland areas. The agricultural area increased 23% in the analyzed period, adding over than 50,000 Km2 of new crops. Most agricultural expansion, and therefore the greatest losses of grassland, concentrates on both sides of Uruguay river (Mesopotamic Pampa and the western portion of Southern and Northern Campos) and the western portion of Inland Pampa. The generated maps open the door for more detailed and spatially explicit modeling of many important aspects of ecosystem functioning, for quantification in the provision of ecosystem services and for more efficient management of natural resources. © 2020 by the authors.