Abstract: Dense non-aqueous phase liquids (DNAPLs) have characteristics such as low solubility, high volatility, and high toxicity, making their remediation particularly challenging. Accurately determining the extent of DNAPLs contamination is a prerequisite for designing an effective remediation plan. The non-invasive induced polarization (IP) method has great potential for application in site characterization of contaminated areas. However, there are many factors affecting the changes in the induced polarization signals of DNAPLs, and obtaining electrical signals through laboratory measurements is often time-consuming, labor-intensive, and subject to significant noise, making it challenging to precisely delineate and characterize the contaminated areas. This study systematically investigates the electrical signal response patterns of DNAPLs-contaminated soils through controlled laboratory batch experiments, and proposes a numerical simulation approach to study the electrical signal response of DNAPLs-contaminated sand columns. The results show that electrolyte conductivity and water saturation both have a negative correlation with phase, while showing a positive correlation with conductivity. This is because an increase in electrolyte conductivity directly leads to an increase in pore water conductivity, which compresses the thickness of the electrical double layer, decreases ion mobility, and reduces polarization intensity. The increase in water saturation enhances the connectivity of pore water in the porous medium, improving conductivity and providing a continuous path for ion migration, thus hindering charge polarization. Both phase and conductivity values increase with the increase in silt content, where cation exchange enhances surface conductivity and polarization effects. Based on COMSOL Multiphysics and MATLAB platforms, a complex conductivity model for unsaturated porous media is proposed, and a series of pore network complex conductivity simulations are conducted to verify and extend the experimental results. The proposed complex conductivity model enables the acquisition of electrical signals from a numerical simulation perspective, and the simulation results under different electrolyte concentrations, water saturations, and soil types are generally consistent with experimental trends. It also addresses the drawbacks of laboratory measurements, such as high noise and long measurement times, and the computed results show better data comparability. This study conducts a multidimensional investigation into DNAPLs contamination through a combination of laboratory measurements and numerical simulations. The results not only provide theoretical and methodological support for the detailed characterization of DNAPLs' existence forms, migration pathways, and spatial distribution characteristics in porous media, but also have significant scientific value and practical engineering guidance for the development of an integrated technology system for DNAPLs contamination site identification, monitoring, and remediation based on non-invasive techniques.