Purpose.To determine the efficacy and safety of topical autologous serum as a treatment of dry eye patients.Methods.A 2-month, prospective, single-masked, placebo-controlled study was conducted in patients with bilateral severe dry eye. One eye was randomized to receive the patient's own serum as a tear substitute, and the fellow eye received unpreserved normal saline solution as a placebo. Subjective symptoms and clinical parameters of dry eye including conjunctival impression cytology were assessed at baseline and 1 week, 1 month, and 2 months after treatment.Results.Twelve dry eye patients were enrolled. Both subjective symptoms (discomfort, foreign-body sensation, dryness, and photophobia), objective signs (fluorescein and rose bengal staining and conjunctival impression cytology) improved significantly in treated eyes compared with baseline. Control eyes also had improvement in symptoms, signs, and rose bengal staining compared with baseline. Neither Schirmer test results nor tear break-up time improved in either group. The means score of all parameters were improved in both groups, and the results of conjunctival impression cytology were better in treated eyes; however, these results are not significantly different. There were no serious adverse effects observed in this study.Conclusions.There was a trend toward improvement in symptoms and signs of dry eye including cytologic changes after application of autologous serum in severe dry eye patients. However, this trend was not statistically significant. A larger scale study is warranted.