This paper presents a multi-physics numerical simulation for detecting void defects in cross-linked polyethylene (XLPE) cable insulation using terahertz (THz) wave technology. Computational electromagnetic models were developed to analyze the interaction between terahertz waves and XLPE material containing void defects of varying sizes and depths. The simulation results demonstrate distinct frequency-dependent electromagnetic responses in the 0.1–10 THz range, with notable differences in electric field distribution, power flow, and reflection coefficients between defective and non-defective samples. Time-domain analysis of Gaussian terahertz pulses revealed characteristic reflection patterns that correlate with void dimensions and locations. The electromagnetic power outflow measured at monitoring points showed systematic dependence on both void size and depth, establishing quantifiable detection thresholds. These findings provide a theoretical foundation for non-destructive evaluation of XLPE cable insulation using terahertz technology and offer in-sights for optimizing practical defect detection systems for power cable quality assurance.