Abstract: In order to accurately identify sediment sources in the gully and hilly zones of the Loess Plateau, this study was used a composite fingerprinting method that was based on two types of fingerprint factors (14 physicochemical properties and 23 biomarker properties (n-alkanes)) with source samples and sediments derived from check dam. The results showed that, neither the individual physicochemical properties nor biomarker properties cannot effectively identify sediment sources. The physicochemical properties were showing the limitation in identifying forest and cropland, and the n-alkanes were showing the limitation in identifying the gully. However, the optimum composite fingerprint consisted of three physicochemical properties and three biomarker properties (CPI, Ca, TP, C₂₀, C₂₉ and Fe) could correctly distinguish 90.5% of the samples. The results showed that in the study area, gully was the main sediment source in this catchment, reaching 60.8%, while the cropland contributed 20.7% of the sediment, and forest and grassland contributed 11.3% and 7.1%, respectively. This research demonstrates that using the composite fingerprint consists of traditional fingerprint properties and biomarkers to identify the sediment source are more accurate, it would be suitable for the area where the geologic variations of a study area are small. The study is meaningful for designing sediment management and soil erosion control strategies in Loess Plateau.