Based on field water quality data from sampling and land use data from interpreting remote sensing imagery, we used correlation analysis and redundancy analysis to study the output relationship between landscape patterns and nitrogen and phosphorus pollution in the Dongjiang River Basin, at different spatial scales and seasons. The results are as follows. (1) In the Dongjiang River Basin, areas of town and water were positively correlated with the nitrogen and phosphorus output, while woodland and grassland areas were negatively correlated with the nitrogen and phosphorus content. (2) In the basin, landscape indices such as Landscape Division Index (DIVISION), Shannon's Diversity Index (SHDI), Interspersion & Juxtaposition Index (IJI), Edge Density (ED), and Patch Density (PD) were positively correlated with river nitrogen and phosphorus content, while Largest Patch Index (LPI), Contagion (CONTAG), and Percentage of Like Adjacencies (PLADJ) were negatively correlated with river nitrogen and phosphorus content. (3) From the perspective of time, the correlation between water quality and landscape pattern was stronger in dry season than that in wet season. From the perspective of spatial scale, the correlation between river water quality and landscape pattern was stronger in catchment scale than that in buffer scale. In addition, nitrogen pollution was more sensitive to landscape patterns than phosphorus pollution.