Land use has obvious influence on surface water quality; thus, it is important to understand the effects of land use patterns on surface water quality. This study explored the relationships between land use patterns and stream nutrient levels, including ammonium-N (NH4 (+)-N), nitrate-N (NO3 (-)-N), total N (TN), dissolved P (DP), and total P (TP) concentrations, in one forest and 12 agricultural catchments in subtropical central China. The results indicated that the TN concentrations ranged between 0.90 and 6.50 mg L(-1) and the TP concentrations ranged between 0.08 and 0.53 mg L(-1), showing that moderate nutrient pollution occurred in the catchments. The proportional areal coverages of forests, paddy fields, tea fields, residential areas, and water had distinct effects on stream nutrient levels. Except for the forest, all studied land use types had a potential to increase stream nutrient levels in the catchments. The land use pattern indices at the landscape level were significantly correlated to N nutrients but rarely correlated to P nutrients in stream water, whereas the influence of the land use pattern indices at the class level on stream water quality differentiated among the land use types and nutrient species. Multiple regression analysis suggested that land use pattern indices at the class level, including patch density (PD), largest patch index (LPI), mean shape index (SHMN), and mean Euclidian nearest neighbor distance (ENNMN), played an intrinsic role in influencing stream nutrient quality, and these four indices explained 35.08 % of the variability of stream nutrient levels in the catchments (p<0.001). Therefore, this research provides useful ideas and insights for land use planners and managers interested in controlling stream nutrient pollution in subtropical central China.