The use of spatial interaction models based on gravity criteria have been widely applied in the study of urban processes. Although the original perception about the similarity between the social phenomena and physics was first suggested by Carey in the middle of the 19th Century, the formal development of models based on gravity algorithms started with Reilly in 1929, and epitomised in the seventies with the work of Wilson. In this paper we first attempt to make a brief overview about the development of models oriented to explain the urban dynamics, and after we explore its possibilities in the simulation of suburban migration, which represents one of the strongest components of the real estate demand on the edges of metropolitan areas, and could be useful in the forecasting of property markets. Finally we offer an empirical implementation of the model for the Metropolitan Area of Barcelona and the central counties of Catalonia.