It's great significance for water quality protection to figure out the effects of landscape composition and configuration in different spatial scales. The Qingyijiang River is the longest branch of lower Yangtze River, the water quality of rivers is influenced by human activities in variety of ways. Taking the Qingyijiang River basin as the research objective, this paper adopted the redundancy analysis (RDA) and Spearman correction analysis to initially explore the impacts of the land utilization and landscape pattern on the water quality, which based on the Google Earth remote sensing data and water quality measurement data. The results are: (1) Among the five buffer scales of 100, 200, 500,1000 and 2000 m established based on sampling point, the landscape composition of the 500 m radius circular buffer zone had the highest interpretation rate of water quality, and the interpretation rates of the low flow period and the high flow period were 46.30% and 43.10%, respectively. (2) In land use types, farmland area and building land area were positively correlated with ammonia nitrogen (NH₄⁺-N), total phosphorus (TP) and electrical conductivity (EC) and produced negative effect on water quality. However, forest land area was positively related to dissolved oxygen (DO), which played a vital role in purifying pollutants. Comprehensive index of the land use was positively related to the polluted indicator, which meant the water quality was increasingly deteriorating in the area with highly intensive human activities. (3) The landscape pattern indexes suggested that patch richness density (PRD) was negatively related to the density of the NH₄⁺-N and TP. The relative indexes were -0.656 and -0.540 respectively, which indicated that the watershed ecosystem became more stable with the increase of PRD. The density of the largest patch index (LPI) was negatively associated with the density of DO in the low flow period, the relative index was -0.656, while area-weighted mean shape index (SHAPE_AM) was positively related to the density of the NH₄⁺-N and TP. The above-mention demonstrated that the growing intensity of human activities would result in the deterioration of water. The relationship between area-weighted mean patch fractal dimension (FRAC_AM) and water quality can not be well described.Through multi-angle analysis, this study revealed the eco-hydrological processes in the Qingyijiang River watershed, and thus is beneficial to land use management and water quality protection, and sustainable use of resources. This may be useful for the relevant functional departments of the government to make decisions.