Modelling framework UrbanSim is applied for analysis and prediction of residential location choice of households in Grand Lyon for understanding the mechanism of spatial distribution of population. The aim of the paper is the evaluation of the predictability of the residential location choice in the analysed urban area with the UrbanSim application. The social component of location choice is mirrored in different variables for different income levels of population as well as in attributes connected with car-ownership. The paper concludes that the predictability can be increased by adding other UrbanSim models to the application.