The increasing frequency and severity of natural and anthropogenic disasters, including those induced by war and climate change, demand innovative tools for monitoring, forecasting, and managing land use change. This paper presents a novel AI-powered Digital Twin (DT) framework tailored for disaster-affected regions, integrating multimodal satellite data, climate reanalysis, and in situ observations. The architecture comprises modular Digital Twin Instances (DTIs), each addressing specific thematic domains, such as vegetation dynamics, land surface temperature, and forest cover dynamics, coordinated through a central Digital Twin Aggregator (DTA). The system supports both rapid and gradual monitoring cycles, enabling timely and scalable assessments. We incorporate recent advances in geospatial foundation models, physics-informed neural networks, and semantic harmonization to address data heterogeneity and scarcity. The framework is demonstrated through pilot applications in Ukraine and Switzerland. In Ukraine, DTIs capture conflict-related cropland losses and forest degradation near the front line, as well as post-flood recovery following the Kakhovka Dam destruction; in Switzerland, annual-scale forest dynamics are assessed, highlighting gradual structural shifts in response to climate and socio-economic drivers. A cognitive user interface further enhances usability by integrating large language models for natural language interaction, improving accessibility for non-technical users. The proposed framework offers a scalable and adaptive approach to land use monitoring, with significant implications for disaster management, environmental recovery, and sustainable development.