OBJECTIVES: To investigate compliance of hypertension patients using modified Theory of Reasoned Action(TRA). METHODS: The data were collected for 7-12 April 1997, by interviewing 190 hypertension patients in Hwachon, Kangwon-do. The analytical techniques employed include contingency table analysis and logit analysis. RESULTS: 15.1% of patients were unaware of the fact that he/she has hypertension and 11.2% did not know that he/she should take drug. 26.8% of patients took drug continuously, 20.1% had drug intermittently, and 53.1% had never have treatment. In the contingency table analysis, several variables were found to be significantly related to patient compliance. They included variables for attitude towards the consequences of taking drugs, normative beliefs, systolic BP at the enrollment, knowledge of how to take hypertensive drugs, variables for general health behavior and experience with having health worker's home visit. The logit analysis was performed by two steps. First step uses experience with drug treatment of hypertension as the dependent variable, and second step uses continuity of treatment. Included in the predictors that are significantly related to the former analysis are subjected norms produced by combining normative beliefs and motivation to comply, knowledge of how to take hypertensive drugs, and opinion about natural recovery of diseases. The only significant determinant of continuous treatment was knowledge of how to take hypertensive drugs. CONCLUSIONS: The results of analysis suggest the usefulness of TRA as a framework for the study of compliance of hypertensive patients. The findings have some practical implication as well. One is that efforts for enhancing compliance should be directed not only patients but also to other persons influencing patient's attitude and behavior. It also suggest that correct understanding of hypertension treatment is essential to perform the appropriate patient role.