<p>Forest and land fires are disastrous phenomena that occur typically during the dry season, leading to severe consequences for the environment, economy, and human health. Sumatra is one of the most susceptible areas to forest and land fires in Indonesia. The research aims to evaluate the potential fire in Sumatra under the global emissions scenarios, RCP26 and RCP85. The potential of fire risk will be estimated using a machine learning-based model, specifically random forest (RF). This RF model is designed to forecast forthcoming fire emissions by taking into account meteorological factors. An empirical fire weather index (FWI) will also be applied to evaluate the fire risk. Both RF and FWI analysis highlight increases in potential fire intensity and frequency due to the chaotic changes in climate conditions, in which the RCP85 scenario is expected to worsen fires by the end of twentyfirst century (2080–2099). The calculation of FWI underscores the significant influence of climate variables such as temperature and relative humidity on determining fire risk, while the RF model demonstrates that alterations in relative humidity and precipitation can potentially affect future patterns of fire. By integrating a machine learning model and FWI calculations, the study offers valuable insights into the prospective fire risks in Sumatra, emphasizing the urgency of implementing proactive measures to mitigate the increasing threat of fires.</p>