Recently, the steel cable used infrastructures are improved with progress in construction technology. In particular, the steel cable in long span bridges such as cable stayed bridges and suspension bridges are critical members which suspend dead load due to the main girders and bridge floor slabs. Damage at cable members can occur in the form of cross-sectional loss caused by corrosion and fracture, and it can lead structural failure due to concentrated stress of the cable. Therefore, nondestructive examination for steel cables is necessary so that the cross-sectional loss can be detected. This study proposes a steel cable health monitoring technique using an elasto-magnetic (E/M) sensor installed in a cable climbing robot. The E/M sensor is applied to detect the cross-sectional loss in this study while it was originally developed for measuring the tensile force in the previous works. To verify the feasibility of the proposed damage detection technique, steel bars which have 4-different diameters were fabricated and the output voltage value was measured at each diameter by the E/M sensor. Optimal input voltage and working point are chosen so that the linearity and resolution of results can ensure through repeated experiments, and then the E/M sensor scans the output voltage of the specimens based on the selected optimal condition. This E/M sensor based monitoring technique will be incorporated into the cable climbing robot in the further research. This proposed approach can be an effective automated tool for steel cable health monitoring.