Modeling land use dynamics is an important component of the landscape level analysis of socio-ecological drivers at the urban-rural interface. We constructed a dataset consisting of 5313 sample points characterizing land cover in 1992 and 2001 across three contiguous counties in West Georgia. We examined two specifications of conditional logit model based the land rent theory: a model of land use allocation, which is similar to models of land use share that utilize aggregated data, and a model of land use change. The land use change model showed better goodness of fit as it controls for cross-sectional effects and allows parameters of explanatory variables to capture temporal effects. The results have implications on land use modeling at watershed or sub-watershed levels.