In an emergency, such as a fire or a terrorist attack, an important aspect is the safe evacuation of people. The unpredictability of the situation makes the task of emergency evacuation one of the main problems. The development of artificial intelligence technologies opens up possible promising ways of developing evacuation management systems. An intelligent evacuation management system can help track the movement and coordinates of people and related critical factors during the evacuation process (for example, slowing down traffic along the paths, blocking the paths, etc.). The study is devoted to the construction of a decision support system for the evacuation of people from buildings based on the results of modeling the development of a fire, intelligent forecasting of the evacuation time in the conditions of the current situation during a fire. The system allows dynamically forming optimal evacuation routes in a changing situation. The proposed results allow us to further move on to the development of an automated intelligent evacuation management system, which will ensure increased safety for evacuees, regardless of evacuation scenarios and the current situation in the building.