Forest fires pose a significant environmental threat, contributing to substantial ecological destruction and economic losses. With the accelerating impacts of climate change, including rising temperatures and prolonged droughts, the frequency and intensity of such fires are on the rise, raising urgent concerns for effective management and response strategies. This study employs advanced remote sensing techniques, specifically utilizing Sentinel-2 and Landsat-8 satellite imagery, to evaluate their efficacy in estimating and predicting wildfire risks. By integrating a diverse set of environmental variables—such as local meteorological conditions, vegetation indices, and topographic features—this research implements machine learning models, notably Random Forest and Extreme Gradient Boosting, to create a comprehensive wildfire risk assessment framework. Additionally, the importance of individual predictors in estimating fire risk, revealing that elevation, slope, aspect, and surface temperature substantially influence the models' predictions. The results indicate that the higher spatial resolution of Sentinel-2 data provides more accurate fire risk estimations than Landsat-8 imagery. This work lays the foundation for improved wildfire management strategies and highlights the integration of satellite data and machine learning as a powerful approach to supporting disaster preparedness and resource allocation in fire-prone regions.