Over the last 20 years, urban micro-simulation models have been the focus of much research. The advent of big data and the solid theoretical base represented by random utility theory, consumer theory and operationalized by discrete choice models seemed to have opened unlimited opportunities for urban micro-simulation. However, initial attempts to replace the traditional aggregated comprehensive urban models with comprehensive micro-simulation models – e.g. ILLUMASS, ILUTE, Oregon and UrbanSim – encountered several methodological obstacles that lowered the overly enthusiastic original expectations. The aim of this paper is to contribute to the understanding of the strengths and weaknesses of micro-simulation modeling generally, and of residential mobility modeling specifically. The following methodological issues are discussed based on a series of experimental microsimulation models of residential mobility applied in the catchment area of the medium-sized town of Tábor in the Czech Republic: micro-data availability, methods of data disaggregation, the multicollinearity of environmental factors and the reliability of highly stochastic models.