Two regional groundwater flow models (Neckar catchment, Germany, 14,000 km2, and Southern Ouémé Basin, Benin, 11,000 km2) were developed within the framework of the integrated management project ‘RIVERTWIN’ (www.rivertwin.org). Both models were evaluated with respect to the question if the chosen modelling approaches (multi-layered finite difference numerical flow modelling, steady state and transient) are appropriate in view of the existing management problems in the catchments, the data availability and the hydrogeological and hydrological conditions in the basins. It is shown that neither the model in the well-investigated, data-rich basin in Western Europe with its highly developed water related infrastructure, nor the model in the hydrogeologically less well-known and less developed basin in Western Africa provide results that are fully applicable to the main regional management tasks. In the case of the Ouémé, the groundwater related problems are foremost of local character and therefore cannot be addressed by regional models in a meaningful way. Data scarcity and complex, unfavourable geological conditions (crystalline rocks, discontinuous aquifers) support the conclusion that numerical 3D groundwater flow models are currently not helpful to manage groundwater related management problems in the Ouémé basin. A better understanding of regional hydrological surface and subsurface processes is required first. Methods for a reliable estimation of groundwater recharge and subsequently groundwater availability were identified as the most urgently needed tool for meaningful groundwater management in view of climatic, demographic and land use change. In the Neckar catchment the results of the analysis are less pronounced; here regional groundwater problems could clearly benefit from a physically based 3D model since the hydrogeological system is strictly stratified with several important aquifers in the vertical sequence. As a general conclusion it can be stated that regional scale groundwater flow modelling concepts seem to be difficult to integrate in management systems and difficult to transfer from one basin to another. This means the question of how to represent the groundwater resources appropriately has to be discussed very thoroughly for any new integrated water resources management problem. It is not possible to give a final recommendation on which modelling concept is the most appropriate one in regional integrated modelling and management. Hence, this article is only intended to provide an in depth discussion of the aspects that need to be considered in the process of choosing appropriate modelling concepts.