Abstract: This paper systematically reviews the research progress on multi-scale failure characteristics and mechanical effects of granular slopes. It aims to reveal the cross-scale correlation mechanisms from microstructural evolution to macroscopic catastrophic behavior, so as to improve the accurate prediction and prevention capacity of geological hazards. A comprehensive analysis method integrating field investigation, physical simulation, microstructural characterization and numerical simulation is adopted to deeply analyze the formation mechanism, movement evolution characteristics and deformation-failure mechanisms of granular slopes in different geographical and climatic regions. From the micro-meso scale, the macroscopic mechanical response and evolution path of soil particle microstructural characteristics (e.g., force chain network, contact cementation) under external environmental actions such as rainfall and earthquake are systematically explained. The multi-scale failure path of granular slopes from micro cementation bond fracture, meso contact network reconstruction to macroscopic strain localization is revealed, and a mechanical analysis framework correlating macro-, meso-, and micro-scale parameters is established. The synergistic mechanism of soil arching effect and force chain network in stability control, as well as the influence laws of key factors such as water content and particle gradation are clarified. The current limitations in cross-scale parameter mapping accuracy, timeliness of dynamic evolution simulation and multi-field coupling mechanism analysis are pointed out. Traditional continuum medium models are difficult to fully characterizing the heterogeneous deformation and dynamic failure characteristics of granular materials. A cross-scale numerical platform integrating deep learning, enhanced computed tomography (CT) and multi-physics field coupling should be developed to realize real-time simulation and prediction of granular catastrophe process. An intelligent early warning system based on multi-source information fusion should be constructed, which can provide effective theoretical support and decision-making basis for geological hazard prevention and mitigation.