Forest fire is one of the major disasters that jeopardize the health of forests, and scientific prediction of forest fire is an important basis for forest fire prevention. In this study, using China's old and new forest fire policies as the dividing line, we divided the historical forest fire data of the Inner Mongolia into four periods, modeled the occurrence of forest fires based on the boosted regression tree (BRT) model, predicted the occurrence of forest fires, and explained the differences in the changes of forest fires and fire risks in different periods. The prediction results showed that: (1) the modeling accuracy area under curve (AUC) for all four periods was greater than 0.94, indicating that the BRT model was able to predict the occurrence of forest fires in the study area better; (2) The daily difference in temperature, daily minimum relative humidity, cumulative precipitation in the previous year's spring control, cumulative precipitation in the previous year's autumn control, the previous year's spring control maximum surface temperature elevation, and the distance of the closest road to the fire point were identified as important drivers affecting the occurrence of forest fires in Inner Mongolia. (3) Changes in forest fire risk levels before and after the implementation of the old and new Forest Fire Prevention Regulations were as follows: from 1981 to March 14, 1988, the medium, high, and very high forest fire risk zones were distributed in the eastern part of Hulunbeier, while from 2009 to 2020 the medium, high, and very high forest fire risk zones were distributed in the southern and central part of Hulunbeier, the southwestern part of Chifeng City, the central Xilingol League and Hohhot, the southern Ulanqab and Baotou, and the eastern Ordos. The study helps to understand the drivers of forest fires and the changes in fire risk levels in the Inner Mongolia under the influence of the Forest Fire Prevention Regulations in different periods, and provides a scientific basis for optimizing forest fire management policies and forecasting.