Turkey may face a lot of losses due to its active geology with an active earthquake zone. After the Kahramanmaraş Earthquake Sequences, it clearly understood the importance of mitigation and preparedness for natural disaster management. This study focuses on the disaster assessment after the earthquake in Gaziantep province and benefits from the advantages of the rapid and grand scale analysis, using Synthetic Aperture Radar (SAR) and optical satellite images. Satellite imagery's extensive coverage and ability to deliver near-real-time data make it an important tool for disaster response. SAR images, in particular, are resistant to unfavorable circumstances such as clouds, smoke, and darkness, making them a significant addition to optical images for damage assessment. The study uses Capella SAR and Maxar satellite images to detect structural damages in four degrees. Multi-modal deep learning models are developed using building footprints to assess damage. Both automatic and manual labeling strategies are used to improve categorization accuracy. We combine SAR and optical data to make use of their distinct information representations through the application of multi-modal learning, which makes it possible to analyze seismic damage more thoroughly and robustly. Also, we evaluate the effectiveness of many deep learning architectures, such as CNN-based models, in identifying traits associated with disasters from diverse satellite data. The findings highlight the usefulness of multimodal learning in post-disaster damage assessment and its potential to enhance seismic damage mapping accuracy. The findings are intended to aid local governments and politicians in implementing more precise disaster response measures.