Objectives The spread of forest fires is a complex and variable process, and the direction and speed of its spread are highly susceptible to the influence of terrain, weather, combustible materials and human activities. Therefore, it is of great significance to improve the real-time accuracy of forest fire burning and spreading in forest fire emergency rescue work.

Methods Based on the research of existing fire spread prediction and system, this paper selected Rothermel model and Van Wagner canopy fire critical ignition model as the core calculation model, chose Huygens principle as the core calculation principle, determined the calculation process of forest fire burn spread prediction, and designed and developed a forest fire burn spread prediction platform. The function of dynamic input of scene data, spread of fire range and real-time prediction of key information of fire site was realized. This paper took real forest fire data as an example, relied on the forest fire burning and spreading prediction platform, which used the forest fire burning and spreading prediction model, simulated and predicted the regional wind field changes and the fire spreading process respectively. Meanwhile iteratively optimized the simulation and prediction results of fire spreading by invoking high-precision remote sensing satellite dynamic scenario data. The feasibility and effectiveness of the model and the function of the platform were verified.

Results The influence of slope and wind direction effect on wind field was analyzed by simulating the wind field in the region. It was found that the wind speed and direction changed little in the open plain area on both sides of the river, the wind speed of the upslope wind was lower, and the wind speed of the downslope wind was higher. At the same time, due to the influence of wind direction and terrain, the wind speed in the valley plain region was higher than that in the broad plain region when the trend was parallel to the wind direction, and it was also higher than when the trend in the valley plain region was not parallel to the wind direction. Through the simulation of the forest fire spread process and scope, it was found that the similarity between the fire spread prediction and the actual fire area had significantly improved, especially as the secondary simulation results were more similar. This verified the feasibility and effectiveness of the forest fire spread prediction model and prediction platform.

Conclusions The forecast platform overcame the limitations of static input and output in the traditional forest fire simulation system, combining scene data to enhance the accuracy of spread prediction. It met the actual needs of fire rescue and provided a robust decision-making support for forest fire prevention emergency command and rescue deployment.