Objective Autoimmune cerebellar ataxias were recently reported to be treatable. However, the proportion of patients with cortical cerebellar atrophy of unknown etiology with autoimmune-associated cerebellar ataxia and the actual effectiveness of immunotherapy in these diseases remain unknown. Methods We measured the level of autoantibodies (including anti-gliadin antibody, anti-glutamic acid decarboxylase (GAD) antibody, and anti-thyroid antibody) in 58 Japanese patients with cerebellar ataxia, excluding those with multiple system atrophy, hereditary spinocerebellar ataxia, cancer, or those who were receiving phenytoin, and the efficacy of immunotherapy was assessed. Results Thirty-one of 58 (53%) patients were positive for anti-GAD antibody, anti-gliadin antibody, or antithyroid antibody. Seven of the 12 anti-gliadin antibody-positive patients, three of the four anti-GAD antibody-positive patients, and three of the six anti-thyroid antibody-positive patients responded well to immunotherapy, indicating that 59% of patients with ataxia-associated antibody-positive cerebellar ataxia undergoing immunotherapy responded well. Conclusion Some patients with cerebellar ataxia have autoimmune conditions and diagnosing autoimmune cerebellar ataxia is therefore an important component in the care of patients with this disease entity.