To elucidate the therapeutic responses to corticosteroids and prognostic factors in the treatment of Graves' ophthalmopathy, 23 patients with Graves' ophthalmopathy were treated with intravenous methylprednisolone 0.5 g daily for 3 days, followed by prednisolone 40 mg daily. Dosage was reduced gradually to 10 mg daily over 4 weeks. The symptoms and signs of the eyes were recorded. Orbital computed tomography and urinary glycosaminoglycan excretion testing were performed before treatment. We observed the effectiveness of treatment at 6 months as the endpoint of this study and classified the response as good or poor. We compared the differences in parameters between these two groups. The results showed that 12 (52%) out of 23 patients had a good response, and three (25%) of those with a good response recovered completely. Corticosteroid therapy dramatically improved lacrimation, reduced soreness and congestion of the eyes and ameliorated exophthalmos, lagophthalmos and extraocular muscle movement. Four (36%) of the 11 patients with a poor response had rebound of symptoms and/or signs after the dosage of prednisolone was decreased. Those with a good response were usually younger, and enlargement of the extraocular muscle was also milder than in those with a poor response. This study demonstrated that corticosteroid therapy is effective in Graves' ophthalmopathy and that extraocular muscle size demonstrated by orbital computed tomography is a prognostic factor of therapy.