<p>The Yarlung Zangbo Great Bend is a dramatic U-shaped bend in the Yarlung Zangbo River, where the river makes a sharp turn around the Namcha Barwa massif. This region is highly prone to landslides due to active geological processes and complex terrain. This study enhances landslide susceptibility prediction by analyzing landslide distribution and its driving factors, considering the region’s geomorphic evolution. A total of 2515 landslides were documented through field surveys, remote sensing, and comprehensive data compilation. Ten key conditioning factors associated with geomorphic processes—lithology, topographic relief, slope, aspect, hypsometric Integral (HI), excess topography, fault density, distance to river, stream power index (SPI), and land surface temperature (LST)—were evaluated for their influence on landslide occurrence. Pearson’s correlation coefficient (PCC), variance inflation factor (VIF), and information gain ratio (IGR) were used to evaluate correlations, multicollinearity, and the contribution of each factor. Advanced machine learning models optimized by the sparrow search algorithm (SSA), including SSA-backpropagation neural network (SSA-BP), SSA-support vector machine (SSA-SVM), and SSA-extreme gradient boosting (SSA-XGBoost), were applied to generate landslide susceptibility maps. The SSA-XGBoost model achieved the highest predictive performance (AUC = 0.930), highlighting its effectiveness in this rapidly evolving geomorphic region. High susceptibility zones were concentrated along steep slopes and deeply incised valleys, particularly near the Yarlung Zangbo, Palong Zangbo, Yigong Zangbo rivers, and in glacier-carved valleys. This study introduces innovative spatial prediction indices and establishes a comprehensive framework for landslide susceptibility assessment, providing scientific support for disaster prevention and mitigation in tectonically active terrains.</p>