Lake ecosystems in arid and semi-arid regions face unprecedented degradation threats driven by climate variability, hydrological fluctuations, and anthropogenic pressures. Despite their ecological significance, systematic multi-temporal geo-hazard assessments of lake margins remain limited, particularly in data-scarce regions. This study presents a novel framework integrating Sentinel-1 Synthetic Aperture Radar (SAR) backscatter metrics and Sentinel-2 optical indices to quantify annual geo-hazard susceptibility and shoreline vulnerability across Khabeki and Uchhali lakes, Soan Valley, Pakistan, from 2016 to 2025. We developed a composite Lake Geo-Hazard Index (LGHI) incorporating terrain conditioning, vegetation vulnerability, bare soil exposure, hydrological forcing, shoreline instability, and SAR-derived moisture stress. Annual LGHI maps reveal persistent high-hazard zones along 35%-42% of Khabeki's eastern shoreline and 28%-38% of Uchhali's northern margin, characterized by steep, unvegetated banks with elevated SAR moisture variability. Correlation analysis confirms strong negative associations between vegetation density and SAR moisture instability, and strong positive associations between water indices and SAR backscatter, validating the physical consistency of multi-sensor integration. Land use/land cover analysis indicates a 12.3% increase in bare ground and 8.7% decline in wetland vegetation within 3 km buffer zones, exacerbating bank destabilization. Our findings reveal progressive shoreline degradation driven by vegetation loss, soil exposure, and amplified moisture stress. The framework provides actionable spatial intelligence for targeted stabilization, ecosystem restoration, and early-warning monitoring in degrading semi-arid lake systems.