Abstract: Objectives: In recent years, advances in Mars exploration missions and high-resolution remote sensing technologies have increasingly revealed diverse geomorphic processes on the Martian surface, including dynamic features such as landslides, water-ice related landforms, and impact craters. As a key dynamic process shaping topography, landslides not only alter local terrain but also potentially hold crucial information about geological environments and climatic evolution history, making them a significant focus in planetary geological research. Mars is characterized by widespread, exceptionally large-scale landslides. Studying these landslides can provide vital clues for understanding Martian environmental conditions—including meteorology, hydrology, topography, and geological structures—and offer essential support for future human missions to Mars. Methods: Although Martian landslides have been extensively studied abroad, systematic research within China remains limited. Leveraging multi-source high-resolution optical remote sensing data and digital elevation models, this study systematically interpreted and constructed a global Martian landslide database. We proposed a novel classification scheme based on kinematic and geomorphic characteristics, uncovered the distribution patterns and movement mechanisms of these landslides, and conducted a comparative analysis of their disparities and analogies with terrestrial landslides. Results: A total of 2,971 Martian landslides were interpreted and categorized into: Sliding Landslides (1 787 sites), exhibiting fluidized movement and long-runout deposits with longitudinal stripes; Slumping Landslides (312 sites), with limited areal extent and short travel distances; Flowing Landslides (50 sites), formed by channelized debris flow with distinct wet-flow features; and Composite Landslides (813 sites), representing transitional or hybrid forms. Conclusions: Landslides are predominantly concentrated in Valles Marineris (about 75%), volcanic provinces, and large impact crater rims. Landslide areas vary dramatically, ranging from approximately 0.077 8 km² to a maximum of 10 351.63 km² (excluding the Olympus Mons aureole), far exceeding typical terrestrial scales. The distribution elevation is primarily concentrated between -5 000-2 000 m, indicating a close relationship between landslide development and specific geomorphic units and topographic conditions on Mars. Geometrically, Sliding-type landslides exhibit the highest estimated length-to-width ratios. Approximately 95% of Martian landslides are classified as 'wide' landslides. The equivalent friction coefficient (H/L) is concentrated between 0.2 and 0.4, with only a minimal fraction of landslides exceeding 1, suggesting that Martian landslides generally possess high mobility and fluidity. Comparisons reveal both differences and similarities between Martian and terrestrial landslides. Martian landslides are immense, often tens to hundreds of times larger than their Earth counterparts. In terms of distribution, both are concentrated in highrelief canyon areas, but their triggering mechanisms differ significantly. Morphologically, Martian landslides display sharper boundaries (including their perimeters and the scarp-deposit distinction) due to the lack of vegetation and limited post-formational modification. Comparative analysis of different landslide types shows: the geomorphic features of Martian sliding landslides (e.g., longitudinal stripes) resemble those of terrestrial glacial and loess landslides; slumping landslides are analogous to large rock avalanches on Earth; flowing landslides share similarities with terrestrial debris flows and rock avalanche-debris flows; and composite landslides are comparable to multistage landslides or "rock avalanche–landslide–debris flow disaster chains" on Earth. These similarities suggest that, despite differing planetary environments, geological processes under gravity follow analogous dynamic mechanisms. The "lack of active external dynamic modification" and "extreme long-term preservation conditions" are identified as the core unique characteristics distinguishing Martian landslides from those on Earth. Study limitations include:(1) Limitations in image resolution may have resulted in omissions or misclassifications within the current landslide inventory, necessitating further verification and supplementation. (2) The absence of thickness parameters precludes accurate volume calculations, thereby constraining in-depth analysis of dynamic mechanisms. Subsequent research should incorporate high-resolution DEMs to estimate landslide thickness and volume. (3) The relationship between flow-type landslides and geological environmental factors has not been systematically examined. Future work should integrate Martian conditions—such as low gravity and minimal atmospheric resistance—and employ methods like numerical modeling to elucidate the underlying physical and kinematic processes. (4) The causative factors for variations in landslide types (e.g., sliding, flowing) within similar steep slope units (e.g., Valles Marineris, impact crater rims) remain unclear. Further investigation is required to clarify the correlations between landslide categories and topographic or geological conditions. (5) The current comparative analysis is restricted to terrestrial landslides, which limits the planetary representativeness of the findings. Future studies should extend comparisons to other planetary bodies, such as Venus and the Moon, to establish a more comprehensive framework for analyzing landslide diversity across planetary systems.