<p>The inverse grading characteristics of granular deposits are commonly observed in landslide dams, influencing the behavior of these deposits. However, the effects of water on this characteristic remain unclear. This study conducted a series of experiments to investigate how water impacts the morphology and particle sorting within granular deposits. Numerical simulations using coupling CFD (Computational Fluid Dynamics) with DEM (Discrete Element Method) were performed to quantitatively analyze the inverse grading characteristics in the deposits and to explore the mechanisms behind these characteristics. The results indicate that the presence of water affects both the morphology and the inverse grading characteristics of the granular deposit due to water-particle interactions. The inverse grading characteristic of the deposit weakens as the water level increases, due to the resistance exerted on the particle motion by the water force. The sliding length of the granular flow also affects the particle sorting by altering the particle velocity upon entering the water, although this effect is less significant than that of the water level. The degree of inverse grading within the deposits can be characterized using the coefficient of variation of particle centroids. This coefficient of variation is mainly affected by the water level, which can decrease from 0.24 to 0.03 as the water level increases from 0 to 23 times the mean particle size. Finally, a model was developed to predict the inverse grading of underwater deposits through multi-parameter regression, considering factors such as water depth, sliding length, and particle size.</p>