Computer-simulated microworlds bridge the gap between the complexity of field investigations and the rigor of laboratory studies. The purpose of this paper is to draw attention to the major methodological issues involved in developing and using computer-simulated microworlds for the psychological study of complex decision-making behavior. These issues comprise flexibility and generality in the microworlds that can be created, adequate psychological validity, automatic trial administration and data logging, capacity to interface simulations of psychological processes, and adequate testing and documentation. Each of these issues is illustrated by indicating how it has been addressed in Fire Chief, a microworld generating program specifically designed for creating realistic, yet controllable, decision-making task environments in the psychology laboratory. Fire Chief (Omodei & Wearing, 1993a) is introduced as both fully portable to other research teams and suitable for use in many different contexts requiring a complex task situation.