Image compression is essential for reducing storage space and transmission bandwidth while maintaining image quality. Conventional compression methods such as JPEG often suffer from visible artifacts and quality degradation at low bit rates. This paper proposes a deep learning-based image compression model using a convolutional neural network (CNN) with an encoder–decoder architecture. The encoder learns compact feature representations of input images, while the decoder reconstructs high-quality images from the compressed features. The model is trained to minimize reconstruction error and preserve important visual details. The proposed approach is evaluated using standard performance metrics, including Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Compression Ratio (CR), and Bits Per Pixel (BPP). Experimental results show that the proposed method provides better image quality and higher compression efficiency than conventional techniques, especially at low bit rates. The proposed model is suitable for multimedia applications, cloud storage, medical imaging, and wireless image transmission where efficient compression and high reconstruction quality are required.