Many of the models designed to support the hydrologic studies associated with urban water resources planning require input parameters that are defined in terms of land cover. Estimating the land cover is a difficult and expensive task and any innovation that can reduce these problems should be of significant value to the water resources planning community. The purpose of the reported investigation was to compare conventional and Landsat-based methods for determining the land cover inputs of hydrologic planning and design models. Comparisons were based on a case study of the Fourmile Run Watershed in Virginia. Results of the study indicated that for planning model studies the Landsat-based approach is highly cost-effective. However, in the design model investigations the errors encountered when the Landsat date are used were unacceptable.