Abstract: The changes in runoff and sediment in the Yellow River is a hot issue in the field of hydrology and river management. The Kuye River located in the Loess Plateau was selected as a typical basin. Based on the observed hydro-meteorological data during 1960-2014, there was a statistically significant increasing trend in annual mean temperature, an insignificant decreasing trend in annual precipitation, and significant decreasing trends in annual runoff and sediment. 1980 and 1999 were two breakpoints of the relationships among runoff, sediment and climatic factors. The ensemble modelling framework of runoff and sediment was constructed by the VIC model and eight sediment models based on machine learning algorithm. The Nash-Sutcliffe efficiency coefficient and correlation coefficient for monthly runoff and sediment simulation were higher than 0.6 and 0.7, respectively, that denoted acceptable performance. Based on the comparison between the reconstructed natural runoff and sediment versus observed values, the attribution of runoff and sediment changes were quantitatively analyzed. The results indicated that the contributions of climate change and human activities to runoff reduction in 1980-2014 were 24%-39% and 61%-76% respectively, and the contributions to sediment reduction were 15%-36% and 64-85%, respectively, compared with that in the natural period from 1960 to 1979. Human activities are the main reason for the decreases in runoff and sediment in the Kuye River basin. The related results could not only improve an innovative methodology to understand the attribution of runoff and sediment changes, but also play an important supporting for the utilization of water resources and conservation of runoff and sediment in the Yellow River basin.