Erosion has been identified as a primary factor in land degradation across varying habitats worldwide. Anthropogenic activities including construction, increased impervious cover, and changes in stream channels alter hydrology and sediment loads. The highly urbanized Los Laureles Canyon Watershed (LLCW) in Tijuana, Mexico crosses the US-Mexico border and discharges high sediment loads into the Tijuana Estuary in San Diego County. LLCW was selected to study how knowledge of erosional processes can be strengthened with geochemical analysis of sediments. Sediment samples (n= 36) were collected from hillslopes to characterize sources, and in sediment traps and the Tijuana River Estuary (TJE). Sediment was analyzed for particle size using a laser particle size analyzer and for geochemistry using X-ray fluorescence (XRF) spectroscopy. Multivariate statistical analyses were paired with sediment fingerprinting techniques to determine the proportion of each source present in the trap and estuary samples. Three sediment source groups were identified based on geochemistry: 1) An unconsolidated fine-textured group with small amounts of cobbles, light-gray color, high concentrations of Na2O, CaO, and MgO, and low concentration of SiO2; 2) A red to brown colored group with high concentrations of FeO, high amounts of cobbles, and a coarser grain size; 3) A cemented group with gray color and coarser grain size. low quantities of cobbles, high concentration of SiO2 and low concentration of FeO . Two sediment source groups were identified based on particle size. One source group has smaller particle sizes, while the second has larger particle sizes. Sediment fingerprinting suggests that 94.2% of the LLCW sediment load is comprised of two sources with distinct qualitative properties (color, cobble presence) and particle sizes. A comparison of source apportionment results with predictions generated by the Annualized Agricultural Non-Point Source (AnnAGNPS) model for LLCW indicates that the AnnAGNPS model overestimated the contribution of non-Lf sediment and could be further constrained by sediment fingerprinting source apportionment. This study contributes knowledge about the degree to which modeling and geochemical analysis can be combined in complex urban border regions to predict sediment sources, with implications for management strategies to reduce cross-border sediment loads.