Nutrient impairment of streams is an acute problem in the southeastern region of Kentucky. The impairment of streams in Kentucky is primarily attributed to waste water treatments plants, failing septic systems and high phosphate laden limestone in the region. In the present context, modeling the nutrient loadings is important for TMDL development and allocation of state and federal funding for improving the water quality. Therefore, nutrients (total phosphorus and total nitrogen) data collected from impaired streams in six watersheds of Southeastern Kentucky are analyzed, and inductive models are developed using two different data-driven techniques. Ten years of monitoring data obtained from several state and federal monitoring agencies (ambient monitoring network of Division of Water, TMDL, STORET database, Watershed Watch) is used. The inductive modeling approaches are used to correlate pollutant loadings in streams with an easily measurable hydrologic parameter/factor. In this study, streamflow is used as the hydrologic parameter for model development. Artificial neural networks (ANN) and nonlinear regression techniques are used to develop inductive models. Modeling is carried out at 8 and 11-digit USGS hydrologic unit code (HUC) watershed units and is a part of PRIDE (Personal Responsibility in Desirable Environment) assessment project for evaluation of water quality. Data requirements, predictive and generalization capabilities of these models are evaluated. Spatial and temporal variations of nutrient loadings in streams along with annual and intra-annual loading statistics are analyzed. Results from the models are used to comment on their suitability of use in water quality management of streams and development of Total Maximum Daily Load (TMDL). Use of inductive modeling approaches to replace highly parameterized process-based modeling approaches under extremely data-poor situations, and their use in devising cost-effective water quality sampling networks is evaluated. Relative contributions of nutrient loadings from different watersheds are used for prioritization of monitoring and allocation of funds within the PRIDE project.