Agent-based systems are being used as decision support systems for solving architectural design problems. Usually in design phase, user behaviour is ignored by the designers. Therefore, after the construction, the users face difficulties in emergency situations and in daily usage. As a result, buildings become insufficient to respond to usersi needs and design goals of the building itself. Before construction, ability of testing the interactions between the building and its users is particularly important to solve the problems in early phases of the design. Hence, to design a building that functions better in certain situations and time loss decreases in the design process. For these reasons, the aim of this study is to develop a model to simulate users in shopping malls.