OBJECTIVE: To determine whether clinical characteristics and signal and morphologic changes on magnetic resonance (MR) images of the spinal cord (SC) are associated with surgical outcomes for cervical spondylotic myelopathy (CSM). PATIENTS AND METHODS: A total of 113 consecutive patients with cervical myelopathy underwent cervical decompression surgery in our hospital from January 2015 to January 2018. All patients with preoperative MR images available for review were recruited for this study. Research data included patient sex, age, duration of symptoms, surgical approach, compression level, preoperative mJOA (modified Japanese Orthopaedic Association) score, postoperative mJOA recovery rate, and complications. Imaging data included signal changes on T2-weighted MRI images (grade and extension on sagittal images, four types of signal changes on axial images according to the Ax-CCM system), SC compression, transverse area of the SC, and compression ratio. The t-test, Mann-Whitney U-test, Kruskal-Wallis H - test, analysis of variance, and regression analysis were used to evaluate the effects of individual predictors on surgical outcomes. RESULTS: The study cohort included 85 males and 27 females with a mean age of 60.92 ± 8.93 years. The mean mJOA score improved from 10.24 ± 1.69 preoperatively to 15.11 ± 2.05 at the final follow-up (p <  0.001). Patients in the poor outcome group were more likely to present with a longer duration of symptoms (p <  0.001) and smaller transverse area of the SC (p <  0.001). Bright T2-weighted high signal changes (T2HSCs), multisegmental high signal changes on sagittal MR images, and fuzzy focal T2HSCs on axial MR images were associated with a poor outcome (p < 0.001, p = 0.005, p <  0.001, respectively). The maximum SC compression and compression ratio were not reliable predictors of surgical outcomes (p =  0.375, p =  0.055, respectively). The result of multivariate stepwise logistic regression showed that a longer duration of symptoms, multisegmental T2HSCs on sagittal MR images and fuzzy focal T2HSCs on axial MR images were significant risk factors of poor outcomes (p < 0.001, p = 0.049, p =  0.016, respectively). CONCLUSION: A longer duration of symptom, multisegmental T2HSCs on sagittal MR images, and fuzzy focal T2HSCs on axial MR images were highly predictive of a poor surgical outcome for CSM. Smaller transverse area of the SC and bright T2HSCs were also associated with the prognosis of CSM.