Objective To observe the influence of edaravone perfusion via the pulmonary artery on postoperative lung tissue and lung function during pulmonary ischemia in deep hypothermic circulatory arrest （DHCA）， and explore its possible mechanism. Methods A total of 24 healthy New Zealand white big-ear rabbits were randomly divided into three groups： （1） control group： DHCA model under cardiopulmonary bypass （CPB） was established； （2）low potassium dextran （LPD）group： LPD solution perfusion via the pulmonary artery after the establishment of DHCA； （3）edaravone group：LPD solution containing edaravone （5 mg/kg） perfusion via the pulmonary artery after the establishment of DHCA. Oxygenation index and lung compliance were observed at the time of baseline condition， recovery of ventilation， 1 hour and 2 hours after recovery of ventilation， and postoperative lung function of the three groups were compared. Malondialdehyde （MDA） and superoxide dismutase （SOD） in pulmonary venous blood were measured. All the rabbits were sacrificed after the operation. HE staining and immunohistochemistry were performed on the lung tissues to investigate lung structure changes and inflammatory reaction. Transmission electron microscopy was used to compare ultrastructural changes of lung.Results There were no statistical difference in oxygenation index， lung compliance， MDA and SOD among the 3 groups under the baseline condition （P＞0.05）. After recovery of ventilation， oxygenation index and lung compliance deteriorated to varying degrees in all 3 groups. Oxygenation index and lung compliance of the control group and LPD group at the time of recovery of ventilation， 1 hour and 2 hours after recovery of ventilation were significantly lower than those of edaravone group （oxygenation index：recovery of ventilation and in control group and edaravone group: 198.25±11.02 mm Hg vs. 244.87±13.05 mm Hg；lung compliance：one hour after recovery ventilation in control group and edaravone group：45.88±1.64 ml/cm H2O vs. 59.75±2.38 ml/cm H2O；P＜0.05）. After CPB removal， MDA levels were increased to varying degrees in all 3 groups. MDA levels of the control group and LPD group at the time of CPB removal， 1 hour and 2 hours after CPB removal were significantly higher than those of edaravone group （P＜0.05）. After CPB removal， SOD levels were decreased to varying degrees in all 3 groups. SOD levels of the control group and LPD group at the time of CPB removal， 1 hour and 2 hours after CPB removal were significantly lower than those of edaravone group （P＜0.05）. HE staining showed clear lung structure， less red blood cell leakage， less inflammatory cell infiltration， and less alveolar fluid accumulation in the edaravone group. Immunohistochemistry showed that integral light density of interleukin 6 （IL-6）in edaravone group was significantly lower than those of the LPD group and control group （14.44±1.75 vs. 20.18±2.22， P＜0.05）. Transmission electron microscopy showed integral basement membrane structure， clear blood gas barrier structure， significantly larger number of type II epithelial cells， abundant but not swollen mitochondria and lamellar bodies in the cytoplasm in the edaravone group， which were destroyed in varying degrees in the LPD group and control group. Conclusion Pulmonary artery perfusion of protective solution in low temperature can significantly reduce lung injury induced by DHCA and CPB. Protective solution containing edaravone in low temperature can better decrease lung injury and protect oxygenation.