Abstract: Forest fires not only cause serious damage to the regional ecological environment, but also lead to huge losses of human life and property, so it is very important to study effective fire monitoring algorithms. In this context, the research was carried out by using the data of FY-4 satellite in China, and the forest fire monitoring algorithm suitable for Liangshan region was determined by comparison. Based on the data of the microwave infrared scanner of Gaofen-4 satellite, the size and location of the fire spot were successfully determined, and the burned areas were extracted, and the most suitable method for Liangshan area was optimized. The research results showed that contextual algorithms had the best performance in forest fire monitoring in the Liangshan Autonomous Prefecture, with the lowest missed detection rate and the ability to accurately detect the location of small fire points. The comprehensive index of the algorithm reached 0.7 or above, providing a scientific reference for forest fire suppression serve in the Liangshan area. By setting the threshold value of infrared scanner data of Gaofen-4 satellite to 330 K, the size and location of fire spots were further determined, which made up for the lack of spatial resolution of FY-4 satellite and provided better technical support for forest fire monitoring. Compared with the gray attenuation maps of different methods based on the data of Gaofen-4 satellite, the attenuation change of infrared band was the most obvious, followed by the global environmental monitoring vegetation index and normalized vegetation index, and the attenuation change of fire zone identification index was not obvious. According to the gray attenuation map in infrared band, the area range of the burned area could be accurately extracted.