Purpose Standard surgical treatment for symptomatic cervical disc disease has been discectomy and fusion, but the use of arthroplasty, designed to preserve motion, has increased, and most studies report clinical outcome in its favor. Few of these trials, however, blinded the patients. We, therefore, conducted the Norwegian Cervical Arthroplasty Trial, and present 2-year clinical outcome after arthroplasty or fusion. Methods This multicenter trial included 136 patients with single-level cervical disc disease. The patients were randomized to arthroplasty or fusion, and blinded to the treatment modality. The surgical team was blinded to randomization until nerve root decompression was completed. Primary outcome was the self-rated Neck Disability Index. Secondary outcomes were the numeric rating scale for pain and quality of life questionnaires Short Form-36 and EuroQol-5Dimension-3 Level. Results There was a significant improvement in the primary and all secondary outcomes from baseline to 2-year follow-up for both arthroplasty and fusion (P \ 0.001), and no observed significant between-group differences at any follow-up times. However, linear mixed model analyses, correcting for baseline values, dropouts and missing data, revealed a difference in Neck Disability Index (P = 0.049), and arm pain (P = 0.027) in favor of fusion at 2 years. The duration of surgery was longer (P \ 0.001), and the frequency of reoperations higher (P = 0.029) with arthroplasty. Conclusion The present study showed excellent clinical results and no significant difference between treatments at any scheduled follow-up. However, the rate of index level reoperations was higher and the duration of surgery longer with arthroplasty.