In this paper we present a model of urban growth and its preliminary application to a case study of the phenomena of residential development in the setting of the eastern periphery of Rome, Italy’s capital city. The modeling approach we use synthesises the two typical paradigms widespread in the community of quantitative urban planning: the traditional one, based on cellular automata (CA), and the (relatively) new one, which is agent-based. In particular, our multi-agent system (MAS) is in-between a reactive MAS, with agents carrying out a two-staged decision processin a complex environment, and a model of statistical physics, since we use populations of agents in order to reduce the number of degrees of freedom of the system. While we explicitly model the consumption of agricultural and undeveloped land due to urban growth, our model may be easily integrated as a socio-economic part into a wider decision support system for environmental planning, e.g. our simulations can produce indicators of environmental impact of the growth of the city: electricity consumption, waste production, etc.