This study aimed to compare the diagnostic efficacy of MRI and PET/CT in lung cancer of mouse with spinal metastasis. 40 healthy Balb/c nude mice were selected. 0.1 ml of human lung cancer cell A549 bacterial suspension was injected by the left ventricle injection method to establish a lung cancer spinal metastasis model, and the abnormal signs of the nude mice were closely observed. When the body weight was reduced by 20%, micro PET/CT imaging and small coil MRI imaging were applied after intraperitoneal injection of thiopental anesthesia in nude mice. After the imaging was completed, the nude mouse was dissected and the spinal tumor was removed. The nature of spinal metastases was diagnosed by the pathology department. 5 nude mice died of abdominal infection, 2 nude mice had no spinal tumors, and the remaining 33 nude mice were successfully modeled. 33 nude mice were confirmed by pathology to have 64 metastatic vertebral lesions, among them, there were 7 cervical vertebrae, 24 thoracic vertebrae, 16 lumbar vertebrae, 6 sacral vertebrae and 11 caudal vertebrae. The sensitivity of MRI in the diagnosis of spinal metastases was 78.13%, and specificity was 56.25%. The sensitivity of PET/CT for the diagnosis of spinal metastases was 92.19%, and specificity was 78.95%. The specificity and positive predictive value of PET/CT for the diagnosis of spinal metastases were not significantly different from those of MRI (P> 0.05). The sensitivity, accuracy and negative predictive values were significantly higher than those of MRI (P< 0.05). PET/CT is superior to MRI in the diagnosis of spinal metastases, and its sensitivity, accuracy and negative predictive values were significantly higher than those of MRI (P< 0.05). It is worthy to be further promoted in clinical practice.