Traffic simulation has been extensively used as a decision support tool for efficient traffic management in daily life. During disasters and incidents, traffic simulation can further help rescue teams to understand the current traffic state and the possible impacts of proposed strategies and then to make proper decisions. At this point, the changes in traffic demand should also be considered in the simulation to reflect the latest traffic state. In this paper, a demand-oriented traffic management system for disasters and incidents with a web-based portal is introduced. In this system, not only the real-time data from loop and wireless detectors but also traffic simulation are used for reconstructing and predicting the traffic state. Furthermore, four traffic demand categories for disasters and incidents are proposed in the system. Some respective demand characteristics are collected according to historical cases, such as departure pattern and people’s perception about evacuation. Such characteristics will be used in the system, since they have a great influence on the simulated traffic state and the effectiveness of applied strategies. Finally, the approach how to generate the respective special demands for disasters and incidents with consideration of the normal traffic demand is proposed as well.