This study investigates the spatial correlation and service flow of supply and demand for water purification ecosystem services at multiple scales (i.e., the Taihu Lake Basin, sub-basin, and county) by quantitatively assessing the supply–demand relationship of nitrogen and phosphorus and introducing the SPANS algorithm to characterize the service flow paths. Through quantitative analysis, the supply–demand relationship between nitrogen and phosphorus was evaluated, and the SPANS algorithm was introduced to characterize the service flow paths. The results show that the water purification ecosystem services in the southwestern region and around Taihu Lake exhibit a good supply–demand balance, while a significant supply–demand deficit is observed in the northern and southeastern regions. Service flow analysis indicates that surplus areas are primarily concentrated in hilly and urbanized central regions, whereas deficit areas are mainly located in non-urban centers. Based on these findings, ecological compensation suggestions are proposed, including dynamic adjustment, differentiated compensation, cross-city collaboration, and guidance of social capital participation, to promote continuous improvement in water quality and sustainable development within the basin.