The purpose of this study were to investigate the prescribed medication contents and drug number at discharge, medication compliance of the elderly after discharge, and find out the factors related to medication compliance of the elderly. The data has been collected from 98 elderly patients who were discharged from medical ward of a university hospital in Seoul, during the period from March 2, 1998 to July 10, 1998. At the begining of interview, subjects were screened for dementia with MMSE-K and they were eliminated from the subjects. Compliance with the prescribed medication regimen was measured subjectively by self-report, and objectively by pill count. Other data has been collected through interviews including Medication Complexity Index (MCI) and demographic data. The data were analyzed by descriptive statistics, t-test. One-way ANOVA, Pearson correlation using SPSS-WIN PC program. The results were as the following : 1) The mean age of subjects was 71.4±5.1. And eighty two point seven percent of subjects received over 6 years of formal education which is much higher level of education than average Korean elderly. The total number of medications prescribed per subject ranged from 1 to 15 with a mean of 5.0(S.D=2.6). The most common drug categories wereantihypertensive(12.1%), digestives/ laxatives (12. 1%), cardiac drug (10.8%), antimicrobial(8.8%), antacids(7.6%), and anticoagulants(7.4%). The mean score of Medication Complexity lndex(MCI) was 14.6(S.D=9.1). The subjects who experienced side-effect after discharge were 35.7%. 2) Pill count compliance were 93.0±10.0% and self-reported compliance were 89.0±10.3%. Ninety four point seven percent of subjects reported they were compliant on the dose of prescribed medication. 3) The reasons of medication non-compliance were ① difficult to fit medication regimen into life style(21.8%), ② no expected effect of the medication (14.3%), ③unwanted side effects (13.6%), ④ forgetfulness(12.2%), and ⑤ no need for the medication by their ownjudgement(10.2%). 4) Employed subjects were more compliant than unemployed. But there was no significant difference in compliance on the basis of other demographic variables. 5) The medication compliance decreased with the increased number of drugs, with higher MCI scores, and with lower perceived health status. The compliance was lower in the group who experienced the adverse effects of medications than who did not. 6) The perceived health status and the monthly cost of medication explained 36% of medication compliance. In conclusion, the medication compliance of the subjects who were recently discharged from hospital was quite high. But it was suggested that nurses should understand risk factors of medication noncompliance and develop nursing interventions to reduce noncompliance for elderly patients after discharge from hospital.