This study utilized comprehensive two-dimensional gas chromatography-time of flight mass spectrometry (GC×GC-TOFMS) to perform a non-targeted qualitative analysis of organic compounds in farmland soils (n = 8) near roads of varying grades in Shanghai. After applying blank filtering and spectral library matching with a similarity threshold in excess of 700, more than 1000 mass spectrum peaks were identified across all samples, with 275 ~ 341 compounds exhibiting an retention index (RI) deviation within ±2% of the library values. Sixty-six organic compounds were found consistently at each sampling location, with the most prevalent being polycyclic aromatic hydrocarbons, phthalates, phenols, thiophenes, and tetrachloroethanes, all of which had notably high signal-to-noise ratios. Fisher ratio (FR) analysis identified 29 compounds with marked differences across road grades, with FR values ranging from 22.6 to 60.3, which were derived primarily from chemical products, biomedicine, and personal care items. These discrepancies were likely attributable to local industrial activities, sewage irrigation practices, and residential emissions. Elemental composition analysis indicated that CHO and CH were the predominant constituents of the detected organic compounds, representing 39.9% ~ 59.1% and 30.2% ~ 45.8% of the total peak area, respectively. Additional classification of anthropogenic contaminants highlighted alkanes, polycyclic aromatic hydrocarbons, and phthalates as the primary pollutants in high concentrations within the roadside farmland soils of Shanghai. This indicated that impacts on roadside farmland soil quality from emissions from traffic sources, such as gasoline and diesel combustion, tire wear, and lubricant evaporation and leakage, cannot be ignored. In summary, the application of comprehensive two-dimensional technology enabled precise qualitative screening of numerous previously overlooked organic contaminants in farmland soils, offering a scientific foundation for future environmental management strategies.