Rapid and precise detection systems are necessary because wild forest fires are a serious hazard to infrastructure, human lives, and ecosystems. The ResNet-50 convolutional neural network (CNN) is used in this work to provide a deep learning-based method for detecting wild forest fires. By utilizing ResNet-50’s potent feature extraction capabilities, the model is optimized to categorize images as “Fire” or “No Fire,” guaranteeing excellent precision and recall. To improve model performance on a variety of image datasets, the system integrates a strong data preprocessing pipeline that includes scaling, normalization, and data augmentation techniques including flipping, rotation, and brightness correction. The model is trained used Adam optimizer and emphasizing early stopping mechanism with learning rate of le-4. The model efficiencies on fire forest dataset which includes satellite, drone, and ground captured images.