Aiming at the problem that the large daily fluctuation of meteorological factors was easy to cause the large change of forest fire weather grade forecast, the forest fire weather index model (referred to as the model) was established by referring to the idea of constructing the forest fire weather forecast model of Brown-Davis scheme, using 1 785 forest fire samples and meteorological observation data in Guangxi from 2016 to 2020. The meteorological observation data was input into the model to calculate the daily forest fire risk meteorological monitoring index. Meanwhile, the intelligent grid forecast data was input into the model to calculate the daily forest fire risk meteorological forecast index. Then the weighted algorithm was used to integrate the daily forest fire risk meteorological monitoring index and the daily forest fire risk meteorological forecast index to obtain the forest fire risk meteorological fusion forecast index. Using 789 fire samples from June 2021 to June 2023, the forest fire weather fusion forecast index was verified. The results show that the prediction accuracy of forest fire danger grade in the next 1 to 7 days after the fusion is from 82% to 56%, which is 1% ~ 3% higher than that before the fusion, and basically met the current application of forest fire weather classification forecast in Guangxi.