In an effort to meet the global growing food demand, nutrient pollutants in runoff have also increased due to intensified agricultural practices. For this reason, stakeholders have tried to shift from conventional agricultural practices to best management practices (BMPs). However, the selection and allocation of agricultural BMPs (Ag-BMPs) at a watershed scale, in practice, is very complex. Optimization approaches for selecting and allocating Ag-BMPs have been used, with limitations on the inclusion of temporal and dynamic spatial aspects. To address this issue, the dissertation’s main objective is to build a spatio-temporal multi-objective optimization modelling framework that provides new insights to improve the selection and allocation of Ag-BMPs in a watershed. To achieve the objective, the optimization framework has been developed, and it allows for incorporation of a greater number of crops and Ag-BMPs scenarios in the optimization model, as well as contemplating the space and time variations to allocate Ag-BMPs. In this framework, the SWAT hydrological model is coupled with the multi-objective optimization algorithm (NSGA-II). Minimization of nitrate (NO3-N) losses and maximization of crop yields at field level were the objective functions. The developed framework and experience with it on the considered case study in Latin America is seen as a useful hydroinformatics tool for supporting management decisions in agriculture.