Non-point source pollution is a cause of major concern within the European Union. This is reflected in increasing public and political focus on a more sustainable use of pesticides, as well as a reduction in diffuse pollution. Climate change will likely to lead to an even more intensive use of pesticides in the future, affecting agriculture in many ways. At the same time, the Water Framework Directive (WFD) and associated EU policies called for a "good" ecological and chemical status to be achieved for water bodies by the end of 2015, currently delayed to 2021-2027 due to a lack of efficiency in policies and timescale of resilience for hydrosystems, especially groundwater systems. Water managers need appropriate and user-friendly tools to design agro-environmental policies. These tools should help them to evaluate the potential impacts of mitigation measures on water resources, more clearly define protected areas, and more efficiently distribute financial incentives to farmers who agree to implement alternative practices. At present, a number of reports point out that water managers do not use appropriate information from monitoring or models to make decisions and set environmental action plans. In this paper, we propose an integrated and collaborative approach to analyzing changes in land use, farming systems, and practices and to assess their effects on agricultural pressure and pesticide transfers to waters. The integrated modeling of agricultural scenario (IMAS) framework draws on a range of data and expert knowledge available within areas where a pesticide action plan can be defined to restore the water quality, French "Grenelle law" catchment areas, French Water Development and Management Plan areas, etc. A so-called "reference scenario" represents the actual soil occupation and pesticide-spraying practices used in both conventional and organic farming. A number of alternative scenarios are then defined in cooperation with stakeholders, including socio-economic conditions for developing alternative agricultural systems or targeting mitigation measures. Our integrated assessment of these scenarios combines the calculation of spatialized environmental indicators with integrated bio-economic modeling. The latter is achieved by a combined use of Soil and Water Assessment Tool (SWAT) modeling with our own purpose-built land use generator module (Generator of Land Use version 2 (GenLU2)) and an economic model developed using General Algebraic Modeling System (GAMS) for cost-effectiveness assessment. This integrated approach is applied to two embedded catchment areas (total area of 360,000 ha) within the Charente river basin (SW France). Our results show that it is possible to differentiate scenarios based on their effectiveness, represented by either evolution of pressure (agro-environmental indicators) or transport into waters (pesticide concentrations). By analyzing the implementation costs borne by farmers, it is possible to identify the most cost-effective scenarios at sub-basin and other aggregated levels (WFD hydrological entities, sensitive areas). Relevant results and indicators are fed into a specifically designed database. Data warehousing is used to provide analyses and outputs at all thematic, temporal, or spatial aggregated levels, defined by the stakeholders (type of crops, herbicides, WFD areas, years), using Spatial On-Line Analytical Processing (SOLAP) tools. The aim of this approach is to allow public policy makers to make more informed and reasoned decisions when managing sensitive areas and/or implementing mitigation measures.