The land use and “source-sink” landscape (SSL) pattern of watershed have complex effects on water quality by influencing the production, transport and absorption of pollutants. Understanding the response relationship of river water quality to watershed SSL is the prerequisite and key to optimize watershed landscape allocation. Based on the Shangwuxi Basin of Lake Qiandao, this paper analyzed the temporal and spatial changes of river water quality indicators such as total nitrogen (TN), total phosphorus (TP), nitrate nitrogen (NO^-₃-N), phosphate (PO^3-₄-P), and permanganate index (COD_Mn). The sub-catchment scale land use, SSL pattern and their effects on river water quality were analyzed. The results showed that water quality of main rivers in the basin gradual increased from upstream to downstream and from tributaries to main rivers. Monthly water quality monitoring on 31 points in non-flood season showed that 46% of the points had TN in Ⅴ- inferior Ⅴ standard, and the average concentration was significantly higher than that in flood season. TP and COD_Mn was mostly lower than class II standard, and was slightly lower than that in flood season. In order to further explore the causes of difference in water quality, this study introduced landscape pattern index and SSL spatial load ratio index (LWLI). Results found that LWLI was significantly positively correlated with water quality indexes such as TN, NO^-₃-N and TP in both flood and non-flood seasons. This indicated that the higher proportion of “source” landscape area in the sub-watershed, the closer to the catchment and the steeper the slope, the greater LWLI values and the higher nitrogen and phosphorus loss. For the flood season, the proportion of tea garden, paddy field and dry land, and landscape pattern indices such as dispersal and juxtaposition index (IJI) and LWLI could explain 46%, 27% and 58% of the spatial variation of TN, TP and COD_Mn concentration in rivers, respectively, while for the non-flood season, they could explain 25%, 46% and 62% of the spatial variation, respectively. From the perspective of SSL pattern, the spatial SSL imbalance existed in Shangwuxi Basin. The “source” landscape was concentrated near the middle and lower reaches of the river course, while the “sink” landscape was mostly distributed on the slopes far away from the river course, which made it difficult for the “sink” landscape to play the role of stream intercept and purification. It is suggested to adjust the land use structure and optimize the landscape configuration. For example, improve the management of “source” landscapes (such as precise fertilization and sewage treatment) and set up “sink” landscapes such as vegetation buffer zones or wetlands in key source areas to promote the interception of pollutants. By introducing the SSL spatial load ratio index, this study deepened the understanding of the relationship between SSL pattern and water quality response, and provided a scientific basis for the improvement of water environment quality in Lake Qiandao Basin.