<p>River confluences, which play a crucial role in navigation and transportation, are subject to continuous sediment deposition processes driven by changes in bed morphology and local hydrodynamic conditions. These natural processes can significantly affect navigation, making it essential to maintain the confluence channel at a minimum depth and width through periodic maintenance dredging. Predicting particle deposition is thus crucial for optimizing dredging operations, ensuring adequate draft for barge traffic, and supporting sustainable transportation. In this study, a three-dimensional Reynolds-Averaged Navier–Stokes (RANS) model coupled with Lagrangian Particle Tracking (LPT) is used to analyze flow turbulence and particle deposition at a natural confluence with a discordant bed. The LPT approach effectively tracks and predicts the behavior and spatial distribution of inertial particles in turbulent flows. The use of such a method appears to be novel within the frameworks of river confluence systems. Polydisperse particles are injected with a wide range of diameters using a log-normal distribution function. Each particle's trajectory is predicted using the particle equation of motion and a stochastic dispersion model, which accounts for particle-turbulence interactions. The coupled RANS and stochastic dispersion models successfully capture the underlying mechanisms at the confluence’s mixing interface. The results reveal the influence of turbulent eddies and gravity on the particle deposition process, highlighting the significant role these mechanisms play in shaping river bed morphology. The evolution of the overall deposition rate is analyzed and aligns well with the theoretically anticipated V-shaped curve variation.</p>