This study investigated soil and groundwater contamination at an in-service oil transportation station in the middle-lower Yellow River Basin, China. Spatial analysis combined with 16S rRNA and ITS sequencing revealed localized heavy metal (Cu, Ni, Cd, Pb) and petroleum hydrocarbon (PHs: 15.0 mg/kg) contamination in the oily sewage treatment area, with vertical migration constrained by silty sand layers. Volatile organic compounds (VOCs) primarily originated from oil tank emissions. Groundwater exhibited hydraulic gradient-driven downstream migration of PHs (0.03–0.04 mg/L) and arsenic (1.1–1.5 μg/L). Indigenous microbial communities exhibited redox-stratified functional differentiation: unclassified Comamonadaceae (Proteobacteria) dominated aerobic zones (monitoring well D5), utilizing nitrate for PHs degradation, while Desulfosporosinus (Firmicutes) mediated sulfate-coupled anaerobic alkane degradation and metal immobilization in anoxic zones (D6). Fungal communities featured Trametes (Basidiomycota), facilitating ligninolytic PAH breakdown via peroxidase secretion. Functional prediction (FAPROTAX/FUNGuild) confirmed a synergistic “fungal preprocessing-bacterial mineralization” mechanism. Microbial metabolic plasticity (e.g., nitrogen respiration, photoautotrophy) enabled adaptation to redox fluctuations. Given the site’s medium-low risk profile, we proposed a tiered management framework: (1) in situ bioremediation that prioritizes indigenous microbes, (2) hierarchical risk zoning, and (3) dynamic monitoring networks. These strategies align with China’s Green Low-Carbon Remediation principles through low-energy microbial technologies. The findings provide a mechanistic basis for balancing industrial operations and ecological health in the Yellow River Basin.