One of the main steps in the solution of the gravimetric boundary value problems of geoid computations is the removal of the terrain effect. This computation requires knowledge of the topography of the Earth in terms of Digital Terrain Models (DTM). During recent years, thanks to satellite radar missions, several DTM’s with different resolutions are computed and presented to the geosciences community. In this paper two main DTM’s namely, GTOPO30 with 30 second resolution, the U.S. Geological Survey's EROS Data Center, and SRTM computed by computed by U.S. NASA within Shuttle Radar Topography Mission, are considered and are applied for local geoid computation without applying Stokes formula at a test region at the Coastal Pars of Iran. As the benchmark the geoidal heights at 50 GPS/Leveling stations within the test region are considered. Results of the computations indicate that SRTM could results in, on average, 30.05 cm better result at the GPS/Leveling point than GTOPO30 at the test region.