Purpose: Magnetic resonance imaging (MRI) is widely used to evaluate intervertebral disc degeneration. Recently, various evaluations of cervical disc degeneration using MRI have been conducted, but there is no gold standard. The purpose of this study was to compare the reproducibilities of previously reported classifications for evaluating cervical disc degeneration by MRI and their associations with clinical symptoms. Participants and methods: A total of 582 subjects underwent conventional MRI of the cervical spine. Disc degeneration was assessed in each intervertebral disc from C2/3 to C7/T1 using five different classifications: Matsumoto’s grading system, Miyazaki’s grading system, Nakashima’s grading system, Jacobs’ grading system, and Suzuki’s grading system. MR images of 30 participants were used, and Cohen’s kappa coefficient of agreement of each classification was calculated for intra-observer and inter-observer reliabilities. These five classifications of disc degeneration and changes of vertebral endplates were measured, and associations with clinical symptoms were assessed. Results: Kappa (κ) values of intra-observer agreement were higher for Jacobs’ classification, whereas those of inter-observer agreement were higher for Nakashima’s and Jacobs’ classifications than for other classifications. The prevalences of neck pain and shoulder stiffness were 27.4% and 41.9%, respectively. There were no associations for any classifications of disc degeneration and Modic types with neck pain or shoulder stiffness. Only the presence of Schmorl’s nodes was associated with neck pain. Conclusion: At present, there is no specific classification for cervical disc degeneration associated with clinical symptoms. Vertebral endplate changes might be associated with clinical symptoms. It may be necessary to create a new classification for better reproducibility of the evaluation of cervical disc degeneration.