Based on the SWAT model, the runoff and sediment processes in three typical tributaries of the Tao River, Zuli River, and Jing River basins in the Yellow River Basin of Gansu Province were simulated. By using double accumulation curves to divide the characteristic periods of runoff and sediment changes, and comparing and analyzing the simulation results of SWAT models in each characteristic period, the contributions of climate and underlying surface factors to the runoff and sediment processes in three typical watersheds were separated. The results show that the SWAT model has high accuracy in simulating runoff and sediment in three typical watersheds, with significant spatial differences in runoff depth and sediment transport modulus among different typical watersheds. The Tao River Basin has more runoff and less sediment, while the Zuli River and Jing River basins have less runoff and more sediment. The slopes of the precipitation-runoff and runoff-sediment double accumulation curves in three typical watersheds show significant changes, with a decrease trend in the runoff coefficient and sediment transport represented by the slopes. From 1980 to 2020, the runoff and sediment in three typical watersheds decreased, and there were differences in the contributions of climate and underlying surface factors to the changes in runoff and sediment during each characteristic period. The influence of climate factors was significant in the early stage, while the contribution of underlying surface factors increased in the later stage. The underlying factors of human activities participating in the transformation are the main reasons for the reduction of runoff and sediment in the Yellow River Basin in Gansu Province. Keywords: runoff and sediment dynamics; SWAT model; double accumulation curves; climate factors; underlying surface factors; the Yellow River Basin of Gansu Province 〖FL