Groundwater managers have often to face regional degradation of the quality of the groundwater under the pressure of diffuse pollution. One of the most problematic and widespread contaminant is certainly nitrate whose concentrations are found in many aquifers to increase over time. The prediction of the future evolution of these trends is of primary importance to manage efficiently the groundwater resource. Unfortunately, the predictive reliability of statistical tools for trend assessment is generally limited. Moreover, classical modeling techniques are often applicable with difficulty at large scale. There is thus a need to develop tools that allow reproducing and predicting these trends. A new modeling technique, the ‘Hybrid Finite Element Mixing Cell’ approach has been developed for large scale modeling purposes. The principle of this method is to fully couple different mathematical and numerical approaches to solve the groundwater flow and solute transport problem. The mathematical and numerical approaches used can thus be adapted to the level of the local hydrogeological knowledge and the amount of available data. In combination with long time series of nitrate concentrations and tritium data, this approach has been used to develop a 3D spatially distributed groundwater flow and solute transport model of the Geer basin (Belgium) of about 400 km2. The model is able to reproduce the temporal evolution of nitrate concentrations and the spatial distribution of nitrate trends. The model has then been used to predict the future evolution of nitrate trends under realistic foreseen scenarios of nitrate load.