This study develops and applies integrated hydrologic-economic optimization modeling for evaluating costs effectiveness of beneficial management practices (BMPs) in agricultural watersheds. The study has three components. Firstly, it defines agriculture as a social-natural system, formulates the system functions, and develops theoretical and empirical frameworks for integrated hydrologic-economic modeling for evaluating BMP cost effectiveness and optimal placement in agricultural watersheds. The integrated modeling contributes to advancing the methodology for watershed-based BMP assessment and improving our understanding on spatial tradeoffs of BMP costs and benefits. Secondly, the study develops a user friendly interface for the integrated modeling as an open-source Whitebox GIS plug-in. The interface facilitates the use of the complex integrated modeling for conducting watershed-based BMP assessment by scientists and conservation managers. Finally, the study empirically applies the integrated modeling to produce new information on the cost effectiveness of representative agricultural BMPs including zero tillage, optimium fertilizer application rates, and red clover as a cover crop in the Gully Creek watershed of southwestern Ontario. The simulation results show that zero tillage reduces the sediment and nitrogen loads by 45% and 43%, respectively, costing the farmers about 2% of the net return. Applying the practice to corn sown after soybeans appears to be profitable. Application of the optimum fertilizer application rates, red clover as a cover crop BMPs, and multiple BMPs are cost-effective in reducing sediment and nutrient loads. Simulation of optimal BMP allocations in the watershed produces several cost-effective BMP policies subject to environmental targets. It also shows that the magnitude of BMP cost depends on the number of available BMP alternatives. The functional relationships between BMP costs and water quality benefits depend on the type of pollutant and the level of environmental target. A piecewise function with steep, moderate, and flat regions better describes the relationships and produces the highest fit. This finding extends our understanding of the BMP cost benefit relationships which are considered to be generally monotonic.