Multi-Agent Systems (MAS) offer a conceptual approach to include multi-actor decision making into models of land use change. Through the simulation based on the MAS, this paper tries to show the application of MAS in the micro scale LUCC. This paper starts with a description of the context of MAS research. Then, based on the BDI (belief, desire, intention) decision-making architecture, it provides a model for the household decision- making. Based upon this framework, Section 4 reports a case study for Mengcha Village of Mizhi County of Shaanxi Province. Section 5 gives the simulated LUCC of 2008, and verifies the result. Section 6 discusses the potentials and drawbacks of the following approach. From our design and implementation of the MAS in a micro scale model, a number of observations and conclusions can be drawn on the implementation and future research directions. (1) The use of BDI decision-making to represent individual households provides a more realistic modeling of the making-decision process. (2) The use of continuous function, not discrete function, which constructs the interaction among households, is more realistic to reflect the effect. (3) This paper attempts to give a quantitative method to analyze the household interaction. And it provides the premise and foundation for researching the communication and learning. (4) The BDI decision-making architecture constructed in this paper helps to accumulate theoretical and practical experience for the interaction research between the micro land use decision-making and the macro land use landscape. Our future research work will focus on the following aspects. (1) The scale issues should be paid attention to the transformation of the household land use decision-making to the collective decision-making. (2) The methods should be explored on the researches into the household decision-making in a longer period of time, so as to build the bridge between long-term LUCC data and short-term household making-decision. (3) The quantitative method and molder, especially the scenario analysis molder, should be researched, which may reflect the interaction among the different household types.