Non-point source pollution is notoriously difficult to assess. A relevant example is ammonia emissions in the Netherlands. Since the mid 1980s the Dutch government has sought to reduce emissions through a wide variety of measures, the effect of which in turn is monitored using modeling techniques. This paper presents the current generation of mineral emission models from agriculture based on micro-simulation of farms in combination with a spatial equilibrium model for the dispersion of manure from excess regions with high livestock intensities within the country to areas with low livestock intensities. The micro-simulation approach retains the richness in the heterogeneity of farm household decision making that are the core cause of the difficulty of assessing non-point source pollution, while using the best available data to track corresponding pollution. Examples are provided that illustrate the strengths of the modeling framework for both pollution monitoring and environmental policy scenario analyses.