<p>To improve river water quality and achieve sustainable utilization of water resources, accurately identifying the sources of river pollutants is crucial. Absolute principal component score-multiple linear regression (APCS-MLR) has been widely applied to identify river pollution sources. However, previous research has typically focused on the distribution of pollution sources at different points in a certain region at the same time or has been limited to monthly water quality data. This study conducted statistical analysis on high-frequency water quality index data collected at 4-h intervals in Yangzhou City, Jiangsu Province, China from 2021 to 2023. By classifying the data according to different quarters and employing principal component analysis (PCA) to extract the main factors, followed by application of the APCS-MLR model, we quantitatively analyzed the sources of river pollution in Yangzhou City across different quarters throughout the study period. The APCS-MLR model demonstrated robust performance with consistently high R² values (e.g., COD and NH₃-N predominantly exceeding 0.70), indicating reliable source apportionment across all quarterly analyses. Results indicate that the overall water quality of rivers in Yangzhou City during the testing period was weakly alkaline, with total nitrogen (TN) significantly exceeding standard limits, identifying as the primary pollutant in Yangzhou's riverine systems. Furthermore, rainfall runoff pollution was consistently identified as the dominant pollution source across all quarters. This study provides valuable insights that can assist policymakers in developing targeted strategies to improve regional water quality.</p>