Background. A model of ischemic injury to the spinal cord was established in pigs by clamping the thoracic descending aorta. The feasibility of monitoring spinal-cord ischemia using the cortical somatosensory evoked potential (CSEP) was evaluated. Outcomes between the simple clamp (SC) group and left-heart bypass (LHB) group were compared. Methods. Twelve mini-pigs were randomly divided into two groups SC group and LHB group randomly. The left intercostal incision was adopted. The thoracic descending aorta was clamped under the left innominate artery outlet. CSEP was recorded to monitor the function of the spinal cord. The behavior score was observed every day after surgery. Mini-pigs were killed on the seventh postoperative day. Ultrastructural changes were observed by transmission electron microscopy. Results. The surgical procedures were successful. Two animals developed incomplete paralysis and the remainder developed paraplegia in the SC group. The animals in the LHB group recovered completely and paralysis/paraplegia was not observed. CSEP measurements showed that the amplitude decreased to <50% of baseline values and prolongation of latency was <10% of baseline values in the SC group but not in the LHB group. Changes in CSEP values were in accordance with changes in behavior score. The ultramicrostructure of spinal-cord tissue under clamping indicated severe damage to tissue layers as well as swelling and damage to mitochondria in the SC group, whereas the morphology was almost normal in the LHB group. Conclusions. These data suggest that during a 30-minute clamping time, left-heart bypass may provide superior spinal protection than simple clamping, and monitoring of spinal-cord ischemia using CSEP was rapid and feasible.