Abstract: The quantitative source tracing of pollution in estuarine river systems serves as a fundamental basis for addressing the challenges of precise pollution control at monitoring sections and achieving refined regional pollution management. Taking the SZ River as a case study, based on hydrological, water quality, and meteorological data, we constructed a high-resolution hydrodynamic-water quality model (IWIND-LRC) by setting open boundaries configured to simulate tidal effects. The model underwent parameter calibration and validation for tidal levels, salinity, and ammonia nitrogen concentrations. Based on the validated model, a numerical source apportionment algorithm was developed to quantify ammonia nitrogen sources at the SZ River estuary. Key findings included: 1) The relative errors between simulated and measured annual averages of ammonia nitrogen during the 2020 calibration and 2021 validation phases remained below 30%. Partial uncertainty updates to model parameters improved agreement between simulations and observations, effectively reproducing the spatial distribution and seasonal variation patterns of ammonia nitrogen. 2) External loads were the primary source of ammonia nitrogen at the SZ River estuary, accounting for 86.06%-88.81%, with major tributaries including the BJ River (20.15%-20.55%), the SW River (15.98%-16.18%), and the WT River (9.53%-9.63%) contributing the most significantly; while sediment internal sources accounted for 10.82%-11.81%. These results provide a scientific foundation for targeted pollution control in the SZ River and offer valuable insights for pollution source apportionment in other riverine systems.