To accurately identify the temporal and spatial characteristics of water quality and pollutant source contributions in the inflow rivers of Taihu Lake， this study conducted water quality evaluation， quantitative analysis of pollutant sources， and verification of targeted pollution control effectiveness based on time-series water quality data from 11 monitoring points. The methods employed included k-nearest neighbor（KNN）， fuzzy comprehensive evaluation， cluster analysis， and the absolute principal component score-multiple linear regression（APCS-MLR） model. Results showed that the water quality of the inflow rivers followed the rule of “more polluted in summer and autumn， and better in winter and spring” temporally， with a pollution peak formed in summer due to agricultural non-point source input and high-temperature algae proliferation. Spatially， the study area was divided into clean areas， transition areas， and pollution risk areas， with nitrogen， phosphorus， and permanganate index as the core driving factors. The APCS-MLR model identified three main pollutant sources： domestic sources， agricultural non-point sources， and shipping-induced sediment resuspension pollution， while the unknown source was dominated by natural background. After targeted remediation at typical points， the proportion of Class Ⅴ water decreased to 0， and the number of pollution-exceeding months reduced from 8 to 1. The multi-model coupled analysis system and “sourcespecific and time-specific” governance plan constructed in this study provide precise and quantitative technical support for the pollution control of the inflow rivers of Taihu Lake.