Expansion of urban area is a well-known phenomenon in developing countries with population growth and the migration from villages to cities being two major factors. Those factors reduce the influence of efforts to limit the cities boundaries. Thus, spatial planners always look for the models that simulate the expansion of urban land-uses, and enable them to prevent unbalanced expansions of cities, and guide the developments to the desired areas. Several models have been developed and evaluated for simulating urban land-use expansions. Although these models are numerous, most of them have focused to simulate urban land-use expansions in sub-urban areas. The regional models that cover wider area are equally important. In this study, a new agent-based model has been developed and implemented to simulate urban land-use expansion in Qazvin and Alborz regions of Qazvin province, which cover 1620 square kilometres. In this model, land-use developers have been treated as computer agents that move in the landscape explicitly, and assess the state of parcels for development. The environment of the model is raster. The agents are categorized based on two scenarios. In the first scenario, all agents are of similar category and in the second scenario the agents are divided into five categories with different objectives. Then, the results of the two scenarios are compared. Due to the spatial essence of the problem, Geographical Information Systems (GIS) were used to prepare the environment of agents’ movement and search, and to aggregate and analyze the results. To evaluate the model, data of year 2005 were used as the input and data of year 2010 were used for checking the results. By calibrating the parameters, the most desired configuration of the model was found in the second scenario, since the results were close to the reality as the Kappa index raised up to 78.17 percent. These results showed that the precision of the model to simulate land-use developments are of considerable quality. Thus, the model is able to detect the area that faced rapid urban expansions. Moreover, a comparison between the results of the two scenarios revealed that dividing the agents into categories with different aims and parameters will improve the outcome of the model. However, it is vitally important to determine the number of the agents in each category as well as their parameters precisely.