This study explores the potential of integrating data science models to enhance the predictive capacity of a theory-guided glacier hydrological model for improved river discharge simulations in the Himalayan basins. By combining data science and physical process models, the study addresses the limitations inherent in each approach.