To help local managers in their decision making for restoring surface water quality, SWAT has been carried out in the context of an integrated modeling process. The concern is focused on nitrate and pesticide concentrations, within two watersheds Né (690 km2) and Boutonne (1300 km2) located in Southwestern France. The agricultural systems and associated practices, heterogeneous in the area, were previously described using typologies (crop rotations * types of soil). At this scale, the SWAT model allows to take into account the diversity and variability of the natural and anthropised systems. But implementing land-use changes on HRUs or sub-basins through the standard GIS interface can be very tedious. Therefore we developed a tool, GenLU2 that builds up, from the typology of soils and agricultural systems, management schedules in a read-in format for the SWAT2009 program database. GenLU2 mainly automates the implementation of land-use and agricultural management changes and assigns them at the HRU scale. For each rotation, GenLU2 decodes it into a succession of single crops. Then it reassembles the management schedule of the rotation and assigns it at the HRU scale. This ability of building up read-in SWAT evolutional scenarios makes GenLU2 an appropriate tool for implementing the SWAT model with complex landuse and agricultural systems. It simplifies the work for assessing the pollution pressure in reference to a baseline scenario, at different scales and time steps. GenLU2 benefits have been assessed by confronting the relevance of the crop rotations modeled by SWAT with those used by agro-environmental indicators.