The Salt Pond region of southern Rhode Island extends from Westerly to Narragansett Bay and forms the natural boundary between the Atlantic Ocean and the shallow, highly permeable freshwater aquifer of the South Coastal Basin. Large inputs of fresh ground water coupled with the low flushing rates to the open ocean make the salt ponds particularly susceptible to eutrophication and bacterial contamination. Ground-water discharge to the salt ponds is an important though poorly quantified source of contaminants, such as dissolved nutrients. A ground-water-flow model was developed and used to delineate the watersheds to the salt ponds, including the areas that contribute ground water directly to the ponds and the areas that contribute ground water to streams that flow into ponds. The model also was used to calculate ground-water fluxes to these coastal areas for long-term average conditions. As part of the modeling analysis, adjustments were made to model input parameters to assess potential uncertainties in model-calculated watershed delineations and in ground-water discharge to the salt ponds. The results of the simulations indicate that flow to the salt ponds is affected primarily by the ease with which water is transmitted through a glacial moraine deposit near the regional ground-water divide, and by the specified recharge rate used in the model simulations. The distribution of the total freshwater flow between direct ground-water discharge and ground-water-derived surface-water (streamflow) discharge to the salt ponds is affected primarily by simulated stream characteristics, including the streambed-aquifer connection and the stream stage. The simulated position of the ground-water divide and, therefore, the model-calculated watershed delineations for the salt ponds, were affected only by changes in the transmissivity of the glacial moraine. Selected changes in other simulated hydraulic parameters had substantial effects on total freshwater discharge and the distribution of direct ground-water discharge and ground-water-derived surface-water (streamflow) discharge to the salt ponds, but still provided a reasonable match to the hydrologic data available for model calibration. To reduce the uncertainty in predictions of watershed areas and ground-water discharge to the salt ponds, additional hydrogeologic data would be required to constrain the model input parameters that have the greatest effect on the simulation results.