Watershed characteristics play a significant role in improving water pollution at watershed scale. Understanding the relationships between watershed characteristics and water quality is of great significance for watershed management. Pharmaceuticals are of increasing concern for their ubiquitous detection and potential threat to aquatic environment. However, little is known about the relationships between watershed characteristics and pharmaceutical pollution in the aquatic environment. To study the relationships between watershed characteristics and pharmaceutical pollution in the aquatic environment, we investigated the occurrence of 29 pharmaceuticals and their quantifying relationships with watershed characteristics in Xiangshan Bay, Eastern China. Pharmaceuticals were extracted by using solid-phase extraction (SPE) and analyzed by ultra-high performance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS). The redundancy analysis and regression modeling were used to explore the correlation between watershed characteristics and pharmaceutical concentrations. The results showed that 22 target pharmaceuticals were detected in the surface water samples, with the detected concentrations ranged from not detected (n.d.) to 220.20 ng/L. Lincomycin (100% of samples), macrolides (100% of samples), antiepileptic drugs (100% of samples), beta-blockers (95% of samples), antidepressant drugs (77% of samples) were the predominant compounds with concentration of 2.36-29.1 ng/L, n.d. -35.8 ng/L, n.d. -37.5 ng/L, n.d.-15.0 ng/L, and n.d.-3.49 ng/L, respectively. Redundancy analysis revealed that the concentrations of pharmaceuticals in the watersheds were positively correlated to topographical features of watersheds. Correlation analysis also showed that the total pharmaceutical concentrations had a significantly negative correlation with mean slope (MS) (P<0.01), and beta-blockers had a significantly negative correlation with slope of area elevation curve (SAEC) (P<0.01). The landscape pattern is also a crucial factor for determining pharmaceutical pollution in the aquatic environment. We have also found that the total pharmaceutical concentrations were negatively correlated with contagion (CONTAG) (P<0.05) and positively correlated with Shannon's evenness index (SHEI) (P<0.05). Furthermore, the area-weighted mean shape index of urban land (IsSHAPE-AM) and the largest patch index of forest (fLPI) had significantly negative correlation with total pharmaceutical concentrations (P<0.05). Overall, the watershed characteristics had significant influences on the pharmaceutical pollution. According to pharmaceutical pollution in the watersheds, the study of the watershed characteristics' correlation to water pollution will help better understanding the influences of the surrounding characteristics on the water quality of surface water at watershed scale. These findings can provide scientific support for improving the water pollution control in surface water at watershed scale from the perspective of sustainable landscape management.