Abstract: Desertification sensitivity assessment is a crucial foundation for desertification control and land management. By effectively identifying areas sensitive to desertification, it provides a scientific basis for targeted prevention and control. Traditional assessment frameworks have predominantly focused on chronic pressures such as climate variability and land use intensity, while insufficient attention has been paid to sudden disturbances including extreme droughts, severe sandstorms, forest and grassland fires, and even social unrest. Against the backdrop of intensified climate change and frequent extreme events, this theoretical gap has become increasingly prominent. This paper systematically reviews the research progress in desertification sensitivity assessment. It first clarifies the conceptual spectrum among desertification sensitivity assessment, desertification assessment, risk assessment, and vulnerability assessment, and reveals the evolution of research methods in this field from static attribute superposition, through human-land system coupling, to data- and process-driven scientific paradigms. Subsequently, taking the Mediterranean Desertification and Land Use (MEDALUS) model, the General Land Desertification Evaluation Model (GLDEM), and the PSR model as examples, it analyzes the current application of three representative assessment systems and highlights their common deficiencies in characterizing sudden disturbances. In response to these theoretical limitations, this paper innovatively proposes the concept of "composite desertification sensitivity", explicitly incorporating the long-neglected dimension of sudden disturbances into the framework of desertification sensitivity assessment. On this basis, a dual-module assessment framework of "chronic baseline + acute disturbance" is further constructed. A formula for calculating the composite sensitivity index is designed, along with a selection scheme for chronic baseline indicators and acute disturbance indicators. The integration pathways of combined weighting methods, threshold trigger mechanisms, and machine learning techniques are also discussed. This framework aims to provide an operational technical solution for desertification sensitivity assessment that integrates both chronic pressures and acute disturbances. Future research should place greater emphasis on technological innovation, data integration, and dynamic analysis to promote the transformation of desertification sensitivity assessment from static diagnosis to dynamic early warning, thereby providing scientific support for the sustainable development of arid zone ecosystems.