Because of many problems associated with traditional procedures for identifying sediment sources, fingerprinting techniques, based on physical, chemical and organic properties of sediment and source materials, are increasingly being used as a valuable and effective alternative approach to assembling such information. In this method, a suitable composite (set) of diagnostic properties and a multivariate mixing model are employed to estimate the relative contribution of sediment sources to sediments transported to basin outlet. In this study, using suitable composites of geochemical elements, radionuclides, organic Carbon, Nitrogen and Phosphorous, capable of discriminating subbasins of the study basin, and a multivariate mixing model were used to determine contributions of that subbasins to sediment yield. The suitable composite fingerprints (elements) were obtained using discriminant analysis. The study basin is known as Margan, located in Pouldasht district, Makoo Township, Western Azarbaijan province. This basin has six subbasins. The suitable composite fingerprints having capability to distinguish the above mentioned subbasins include: Tl and Ce for the subbasins of 1 and 2, W, V, and Pb for the subbasins of 3 and 4, and Ni, Mn and Ce for the subbasins of 5 and 6. Mean contributions from the subbasins of 1, 2, 3, 4, 5 and 6 were to be estimated as 89.14, 10.86, 20.86, 79.14, 67.10 and 32.90% respectively. Low mean absolute errors (less than 11%) show high degree of correspondence between measured and predicted properties. High model efficiencies (greater than 0.99) confirm the goodness of fit of the mixing models. Also it is argued that fingerprinting estimates for sediment sources are consistent with field observations. Then although a number of limitations must be recognized, the fingerprinting approach to source ascription has high efficiency to determine relative importance of sediment sources (subbasins) in the study basin.