This study investigates the food web structure of the Xinlicheng Reservoir, a drinking water source of critical importance in Changchun, China, by employing stable isotope analysis (SIA) to quantify the contribution ratios of four basal dietary sources—phytoplankton, zooplankton, sediment organic matter, and particulate organic matter (POM)—to the diets of two key filter-feeding fish species, Hypophthalm ichthys molitrix and Aristichthys nobilis. The analysis reveals that phytoplankton is the dominant dietary source for both species, contributing 32.08% and 34.06%, respectively, whereas the POM contribution is discernably lower (13.25%). The average trophic level of the fish assemblage in Xinlicheng Reservoir is 3.03, while the trophic levels of the two filter-feeding species lie between 3.00 and 3.50. Furthermore, a random forest model was used to identify key environmental drivers of isotopic variations in these basal dietary sources, highlighting the significant role of pH, total nitrogen (TN), chloride (Cl−), calcium (Ca2+), phosphorus (TP), and silicate (SiO44−) in influencing carbon and nitrogen isotopic ratios. These findings provide critical insights to optimize biomanipulation strategies aimed at improving water quality in drinking water reservoirs by enhancing our understanding of the environmental factors that govern trophic interactions and broader food web dynamics.