Abstract: The characterization of three important hydrological features, drainage networks, ridge networks and watersheds, is essential in the study of the geomorphological organization of a given terrain. In this study, mathematical thinning based algorithms to extract these three hydrological features from Digital Elevation Models (DEMs) are developed. First, the fundamental mathematical morphological operators, in particular morphological thinning based operators, are discussed. Drainage networks, ridge networks and watersheds are extracted using skeletonization by morphological thinning, exoskeletonization by morphological thinning and skeletonization by influence zone, respectively. The effectiveness of the proposed algorithms is tested by implementing them on a simulated DEM and the photogrammetrically generated DEM of Lake Mary. The proposed algorithms are able to operate effectively on flat areas in DEMs and produce complete and connected outputs. The accuracy of the extracted hydrological features is validated by gauging their conformity with Hack`s law. A log-log plot of the length of the longest stream in each extracted catchment against the corresponding catchment area is drawn. A power law relationship is observed between the two parameters. This power law relationship has a computed scaling exponent of 0.73, which is slightly higher than the standard value of 0.60 provided by Hack`s law. This deviation indicates that raw extraction of hydrological features from DEMs is not sufficient to produce highly accurate hydrological features due to errors in DEMs and the results obtained need to be complemented with data from other GIS data captures, such as ground survey maps and Landsat Thematic Mapper images.