Internal defects such as cracks and pores in prebaked anodes severely affect aluminum electrolysis efficiency and energy consumption. Existing detection methods face an irreconcilable contradiction among speed, cost, and accuracy. To address this issue, this paper proposes a current fixed-point excitation detection technique, which inverses local resistivity anomalies through multi-point potential difference measurements, enabling rapid identification of internal defects in anodes. In this study, a numerical model of electrical conduction in prebaked anodes is established using COMSOL Multiphysics 6.0 to systematically analyze the influence of defect geometric dimensions on the surface potential difference distribution. The effectiveness of the method is verified through validation experiments on anodes with prefabricated defects. The results show that defect width has the most significant effect on potential difference; when the width increases from 100 mm to 300 mm, the sheet resistivity of the defect region rises from $50.95 \mathsf{\mu }\mathsf{\Omega }·\mathrm{m}$ to $125.1 \mathsf{\mu }\mathsf{\Omega }·\mathrm{m}$. Under dual-defect conditions, the inter-defect sheet potential difference is lower than the defect-free baseline, exhibiting a “valley” characteristic. The relative error between experimental and simulated potential differences in the middle region of the anode ranges from 3.24% to 9.28%, and significant potential difference peaks appear in all defective sheets. This method provides a feasible technical approach and laboratory-based validation foundation for the online nondestructive testing of internal defects in prebaked anodes, and clearly delineates the developmental stage of the present study.