Objective: To study the efficacy of topical methylprednisolone compared with placebo in patients with Stevens-Johnson syndrome (SJS) and/or toxic epidermal necrolysis (TEN).Methods: Thirty-six patients were prospectively studied at Siriraj Hospital. The clinical severity score for ocular involvement was classified as mild, moderate and severe. In 12 patients with moderate severity, 1% methylprednisolone eye drop was added four times daily in one eye and 0.5% normal saline as placebo in another eye for one week. In 2 severe cases, these medications were added every 2 hours till bedtime for one week and four times daily for another week.Results: The average age of 14 patients was 50 ± 16.2 years (±SD), with female preponderance (12). Common precipitating factors were carbamazepine, nevirapine, phenytoin and dimenhydrinate. The common associated diseases were HIV. Five patients were not given systemic corticosteroid due to infection, and minimal skin involvement. The total severity score change in each group was the same at one month. In an average of 7.6 months follow up, most patients recovered normal appearance except for two eyes of the placebo group and one eye of the methylprednisolone group. The percentage of dry eye in the Schirmer test and the rose bengal score in the topical methylprednisolone eyes were not significantly less than in the control eyes. However, less tear production was statistically significant in patients with systemic corticosteroid.Conclusion: Topical corticosteroid might be used cautiously in early stages of SJS or TEN.