Wildfires are a major natural hazard that can destroy ecosystems and homes. Quickly and accurately identifying burned areas is essential for emergency response and recovery planning. In this study, we tested two common machine learning models—Random Forest (RF) and Support Vector Machine (SVM)—to see how well they can map burned areas using satellite images. We focused on two large wildfires that occurred in early 2025 in California: the Palisades and Eaton fires. We used data from two types of satellites: MODIS, which has coarse resolution but frequent observations, and Landsat 8, which provides finer resolution but less frequent images. For both datasets, we created balanced training samples and evaluated model performance using accuracy, precision, recall, and F1‑score. The results showed that RF consistently performed better than SVM, reaching 99.5% accuracy with MODIS and 95.3% with Landsat 8. SVM performed much worse, especially with the higher‑resolution Landsat data. Our findings suggest that RF is more reliable for burned area mapping, MODIS is better for rapid large‑scale fire detection, and Landsat is more useful for detailed post‑fire damage assessment. Future research could explore other satellite data sources and advanced machine learning or deep learning models to further improve accuracy and generalizability.