The current paper aims to provide a regional-level (Development Regions) assessment of the past changes in the agricultural land use pattern during the 1990-2006 period and to simulate future changes (2007-2050) in order to detect the main potential agricultural flows and their regional differences. The simulation was carried out through the CLUE-S model (the Conversion of Land Use and its Effects at Small regional extent) using CORINE Land Cover (CLC) database and several biophysical and socio-economic explanatory variables associated with the current land use/cover pattern. Because of the political and socio-economic changes that took place after 1990 and their relevance for the resulted spatial transformations in land use/cover pattern, two scenarios based on the annual change rate of the 1990-2000 and 2000-2006 were used in order to explore potential future land use/cover changes. Thus, the predicted maps indicate significant changes in the agricultural land use pattern mainly in relation to the local-level conversion processes. Two change flows stand out, i.e. intensification and extensification of agriculture. The first flow is more likely to occur in the Banat and Criana Plains and Hills, the Transylvanian Tableland and west of the Romanian Plain, while the second will mainly occur in the Moldavian Plateau, the Dobrogea Plateau, the central part of the Romanian Plain and the Subcarpathians. Furthermore, a significant increase in the agricultural lands related to forest losses and decrease related to forest gains are expected in the plain regions, Transylvanian Tableland, Getic Piedmont, Moldavian Plateau and Subcarpathians. The results of the current research provide important information on the future spatial distribution of the agricultural land use classes in order to improve the general understanding on the causes and consequences of change. Therewith, it can be utilized as reference study for further sustainable land use strategies and policy-making to ensure the effective management and use of agricultural land resources.