Reforestation in arid regions is critical for combating desertification, restoring ecosystems, and mitigating climate change. Leveraging AI models tailored for such environments offers a promising solution to monitor vegetation health and predict survival rates effectively. This study focuses on designing and implementing a Convolutional Neural Network (CNN)-based model using Sentinel-2 satellite data to support vegetation health monitoring and survival prediction. The developed model achieves state-of-the-art performance by utilizing vegetation indices such as NDVI and EVI and incorporating transfer learning techniques. Validation against field data and benchmarking with traditional models demonstrate its scalability and applicability for reforestation initiatives such as the Saudi Green Initiative (SGI). This is critical due to increasing desertification and urgent global climate targets.