Reliable assessments of large mammal population sizes are crucial for the management of protected areas. We tested feasibility of foot surveys for population assessments of large mammals in western Tanzanian woodland, comparing estimates of herbivore densities from line-transect data from a National Park with those from an adjacent Game Reserve (GR). We used a Geographic Information System (GIS) and Global Positioning System–supported field design, consisting of systematically distributed closed-circuit transects, and recorded sighting distances and angles. Total survey effort was 1,032 km, conducted within the dry season. We fitted detection functions to distance data with the help of DISTANCE 4.1, using the 3 habitat categories woodland, grassland, and swamp as covariates for detection probability. We found estimates of density and abundance to be reliable for 19 out of 20 larger mammalian herbivores and found significant differences in density between the Park and the GR for 5 species, of which 4 had a higher density in the Park and one had a higher density in the GR. Our results show that, using GIS support and modern navigation methods, foot-transect surveys can be effective in providing accurate data on woodland herbivore populations even in large study areas.