Cellular Automata are common approach to simulate urban expansion and different implementations based on them have done for modeling urban growth in recent years. Scale of model is one of the important properties that are effective on the results of these models in urban expansion area. In this study, the effect of changing scale was considered based on results of modified continuous cellular automata. At the first step, fundamentals and basic components of proposed model were represented and uncertain effects of changing scale in cells were demonstrated. Furthermore, the proposed model was implemented on a real city. The study area is the Isfahan city and land cover data for modeling of this area were derived from satellite images using supervised classification methods. Scale change experiments were done based on two different strategies in this study. In the first strategy the growing factor of cells was constant in experiment but the value of this factor was changed according to cell size during the next experiment. The estimations of error were done by the Root Mean Square error. In the first strategy, the best simulation results were given at phase with the cell size of 68.5 meters and the maximum deviation between the upper and lower error based on this index was 0.0035 meters. In the later strategy, the best simulation results were given at phase with the cell size of 63.5 meters and the maximum deviation between the upper and lower error was 0.1055 meters.