Hydrological processes in plateau mountainous basins are governed by steep vertical climatic gradients and complex topography, resulting in pronounced heterogeneity in time and space that challenges the application of traditional hydrological models. Distributed hydrological models, leveraging their strengths in grid-based spatial discretization and physically based runoff generation mechanisms, are capable of quantitatively simulating land–atmosphere interactions and hydrological responses. Consequently, they are widely employed for hydrological modeling and forecasting, particularly streamflow, in these plateau mountainous basins. This paper reviews the progress in the application of distributed hydrological models to plateau mountainous basins. At the model structure level, the improvement and multi-process integration of modules such as permafrost hydrothermal coupling and glacier dynamic processes are realized. The grid products such as multi-source remote sensing and reanalysis are integrated with ground observation, which significantly improves the accuracy and spatial-temporal coverage of input data. Parameter determination and calibration strategies, including parameter regionalization functions (parameter transfer functions) and remote sensing-constrained calibration techniques are developed to improve the distributed hydrological simulation in data-scarce or ungauged basins. Furthermore, this paper reviews the key achievements of distributed hydrological models in simulating rainfall–runoff processes, glacier-snowmelt dynamics, soil erosion assessment, and hydrological responses to climate and land-use changes. Despite solid progress, current research reveals persistent challenges in hydrological modeling for plateau mountainous basins, such as inadequate representation of glacier dynamics, limited coupling of permafrost-vegetation-hydrology processes, and parameter uncertainty in high-elevation cold regions. Future research should advance the mechanistic understanding of coupled water–energy–carbon cycles, develop machine learning-assisted (or physically based) self-adaptive grid-based parameterization schemes, and establish integrated multi-model and multi-source data frameworks. These advancements are crucial for further improving the predictive capability, accuracy, and reliability of distributed hydrological models in complex plateau mountain environments, thereby underpinning ecological security and sustainable water resource management strategies.