We have proposed and implemented a system that supports evacuation after a large-scale disaster. When a large-scale disaster such as earth-quake or conflagration occurs, it may not be possible to pursue predefined evacuation route due to collapsed buildings or fire. The refugees have to select an optimal evacuation route according to circumstances. In such situations, however, it is almost impossible for refugees to grasp the precise circumstance and find the correct evacuation route. In order to mitigate this situation we have proposed a system based on smart phones and server/client system, and implemented it. We make the server side perform the basic processing for evaluating the dynamic situation based on the information collected from refugees’ smartphones using crowdsourcing technique so that the system configuration is flexible. The evaluation is performed based on the idea of the ant colony optimization (ACO) algorithm on the server side. We have implemented the client side of the evacuation route guiding system on both Android OS and iOS, and the server side on Linux system. We have achieved to construct a practical system applicable for real world assuming network infrastructure is intact.