Water quality poses a serious challenge to a community's well-being, whether that of natural or human ecosystem communities. Nonpoint source pollution is the main cause of water quality problems in the USA since point sources have been remediated after the Clean Water Act was instituted in 1972. Land use policies play an important role to guide land development and subsequently influencing water quality. Diverse water use groups, including interested and affected parties (residents in specialist groups and residents in general), technical specialists (planners and scientists), and executive decision makers (water managers and elected officials) all have stakes at risk and roles to play concerning water quality improvement. A comprehensive set of tools to predict the impact of nonpoint source pollution are needed to inform about the relationship between land use policies and water quality. Stakeholder involvement is needed in watershed management to foster credible decision making in land use policy development and water quality improvement as part of watershed management. The research goal focuses on how land use affects nonpoint source pollution and how nonpoint source pollution affects water quality in the Green-Duwamish Watershed of Puget Sound basin, while considering the potential for stakeholder involvement in water quality management. The research achieves a goal for developing a Geodesign Framework for Water Quality Improvement (GeoFWQI). The framework makes use of a nonpoint source pollution model called OpenNSPECT to simulate total nitrogen in the Green-Duwamish Watershed. A land use change modeling software called ENVISION is used to investigate the impact of total nitrogen from three land use policy scenarios: Status Quo (SQ), Managed Growth (MG), and Unconstrained Growth (UG) for the year 2040. The results of simulating pollution runoff in the OpenNSPECT model revealed that the MG scenario has the least contribution of total nitrogen and the UG scenario has the most contribution of total nitrogen across the watershed. Agriculture areas contribute the most to total nitrogen concentration which are mostly located in the middle elevations of the WRIA9; while forest areas contribute the least to total nitrogen concentration which are mostly located in the upper elevations of the watershed. The GeoFWQI was tested by implementing a role play modeling activity in the GEOG 514 GIS Problem Solving course within the Master of Geographic Information Systems for Sustainability Management program at the University of Washington. The results from the role play modeling activity produced several findings. Public participation is perhaps not appropriate for all stages of decision making. Public involvement requires time and money that might or might not be available. There appears to be conflict between land owners and policy makers. Research about color map design related to stakeholder interpretations is needed to better understand how water quality is understood. The GeoFWQI that organizes how stakeholders can contribute in every phase of the water quality modeling requires further investigation. This research contributes to using a land-use/cover model together with results from a nonpoint source pollution model to create a social-ecological system for improving water quality management decision making. The application to the Green-Duwamish watershed in Puget Sound provided a complex context for evaluation. The GeoFWQI offers a vehicle to drive the integration of the biophysical and social sciences to better understand policy makers and public stakeholders in what is occurring at the watershed scale.