Sediment accumulation in the upper Fenhe River basin threatens reservoir storage and ecological security, reflecting the broader soil erosion challenges across the Loess Plateau. To address the difficulty of tracing sources for the highly homogeneous geochemical conditions in this basin, this study integrated End-Member Mixing Analysis (EMMA) with a Sediment Fingerprinting (SF) analysis. A method was developed to combine the two analysis to determine the sources of the sediment deposited at samping sites throughout the basin. The study found that sediments were primarily sourced from proximal tributaries, with Jianhe, Lanhe, Dongnian, and Xinian Rivers contributing 91.6 %. The SF approach demonstrates higher sensitivity in detecting weak signals and minor sources (< 1 %) and exhibits stronger robustness under homogeneous conditions, whereas EMMA provides greater efficiency in rapidly attributing sediments to dominant sources, underscoring the complementary strengths of the two approaches. Sediment generation is driven by the coupling of loess parent material and anthropogenic disturbance, with mining acting as a key amplification factor and sloping farmland having limited impact. Based on these findings, a zoned management framework targeting the major sediment-contributing tributaries-"Source Interception-Process Regulation-Terminal Stabilization"- is proposed to strengthen full-chain management of fine sediment. This study identifies critical sediment-producing areas and controlling factors, clarifies management priorities, and provides a scientific basis for shifting sediment control in the upper Fenhe River basin from broad remediation to precise regulation, offering practical guidance for ecologically fragile regions of the Loess Plateau.(c) 2026 International Research and Training Center on Erosion and Sedimentation and China Water and Power Press. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/).