Image fusion is a technique wherein data or multiple images from distinct sources are merged to generate a unified composite image that incorporates information from each of the source images. The goal of image fusion is to enhance the quality and information content of the resulting image by merging complementary details from the source images. This can prove to be exceptionally beneficial in a multitude of domains, encompassing computer vision, medical imaging, remote sensing, and more. Convolutional Neural Network (CNN), a form of artificial neural network specifically engineered for the analysis of structured grid data (e.g., image processing), is implemented in this paper. It contains convolutional layers, pooling layers, activation functions, and fully connected layers. CNN has become the standard for image classification, object detection, facial recognition, image segmentation, etc. The ability of CNN to automatically learn features from raw data makes it highly versatile and capable of generalizing well to new, unseen data. Because of CNN feature extraction, hierarchical representation, adaptability, effective handling of spatial data, semantic understanding, transfer learning, and the improvement of quality and informativeness are improved. The expected outcome of this project includes improved image fusion results, demonstrating the effectiveness of CNN-based approaches in this domain.