Changes in habitat are often a major influence on species distribution and even survival. Yet predicting habitat often requires detailed field data that are difficult to acquire, especially on private lands. Therefore, we have developed a model that builds on extensive data that are available from public lands and extends them to surrounding private lands. This model is applied for a five-county region in Georgia to predict habitats for the gopher tortoise (Gopherus polyphemus), based on analysis of documented locations of gopher tortoise burrows at the Fort Benning military installation in west central Georgia. Burrow associations with land cover, soil, topography and water observed within the military installation were analyzed with binary logistic regression. This analysis helped generate a probability map for the occurrence of gopher tortoise burrows in the five-county region surrounding Fort Benning. Ground visits were made to test the accuracy of the model in predicting gopher tortoise habitat. The results showed that information on land cover, soils, and distances to streams and roads can be used to predict gopher tortoise burrows. This approach can be used to better understand and effectively carry out gopher tortoise habitat restoration and preservation activities.