Forest fires are uncontrolled burning of vegetation in forest territory. Its danger lies in its rapid and spontaneous spread, which is difficult to combat, and the consequences are long-term restoration of flora and fauna in the territory where it happened, damage to the ecological, economic, material condition of the terri-tory of the region, physical and psychological health of people. At the same time, the prevention and elimina-tion of forest fires is very expensive. At the same time, an important task is to quickly and effectively eliminate the fire, planned so that the costs were minimal. Depending on the type of forest fire, it is important to make timely decisions on the liquidation of such natural disasters. Such solutions require considerable time, analysis of a large amount of information and can not always be called optimal. Thus, it is urgent to create methods and tools to improve the quality of firefighting forces and facilities. At the same time, the efficiency of operational control under conditions of considerable uncertainty depends not only on the reliability of the information, but also on its qualitative analysis, and in conditions of limited time. The article deals with the mathematical support for more effective decision-making in the elimi-nation of forest fires. As a tool of fuzzy modeling was selected based on the theory of fuzzy sets, fuzzy logic. The substantiation of this approach is presented. Briefly outlines the main provisions of the method of fuzzy modeling. As a control function were chosen pace of barrage bands at different stages of fire fighting. This control is formed as a function of measurement known area of fire and its propagation velocity. The effi-ciency and effectiveness of the algorithm of forming control function were confirmed by numerical simula-tions. In modeling, to simulate a real situation, the fire pattern in the form of non-linear differential equations, developed by the author of this article was used.