In Korea, sudden deaths of middle-aged and older male workers who are the responsible persons at home as well as workplace, have aroused social concern. Besides, U. N. has reported recently that mortality of Korea male in 40-50's was one of the highest among newly developed countries in 1992. Not much is, however, known about the nature of the work contributing to the mortality of different groups of workers. Therefore, this study was done to examine mortality of public servants, comprising about 5% of all the employed in Korea, according to their job titles and grades. The datas of 1753 official deaths, comprised 323 (only disease-oriented deaths) applicants of survivors' compensation, were used to examine age-adjusted Standardized Mortality Ratios(SMRs) according to their job titles, grades, tenures and cause of deaths. Controlling age and sex difference was conducted using 26,950,481 general population, 95,340 general deaths and 864,560 working public servants. All the groups were aged 20-64, who were being observed January - December in 1993, at the same time. Results and discussions are as follows. 1. SMRs standardized by general population was significantly low (SMR 44.9 CI 42.8-52.7) for all job titlses. Of public servants, 90.0% was graduated from high school, although 17.5% in general population. The distinction of social status such a education may produce a strong healthy worker effect. Besides, SMRs for different tenure groups showed a steady increase as tenure increases. This suggests that the magnitude of healthy-worker effect may be greater with increasing tenures. 2. SMRs standardized by own public servants was significantly elevated for work-men(SMR 121.0, CI 110.2-132.6) in solitue. When SMRs for different grade, of work-men was examined, 9th(SMR 124.2, CI 104.4-146.7) and 10th(SMR 137.9 CI 120.8-156.8)grade, lower grade in workmen, showed significantly elevated SMRs. Of workmen, 57.0% were graduated from high school and 50.1% in 9th grade, as well. These mean that low economic states made up social class, education may increase mortality rate. 3. Of SMRs according to all causes of death, only policemen on 'cause of death related hypertensive disease'(SMR 282.5, CI 121.6-556.7) was significantly high except for' cause of death related other signs, symptoms and ill-defined conditions'. 4. When SMRs on cause of death related hypertensive disease for different grade of policemen was examined, senior policemen(SMR 241.9), in charge of the front service, showed elevated SMR, in spite of statistical no significance. Especially, the working hour of senior policemen is quite long and also the work schedule is even more irregular for policemen. The results of this study showed that mortality for different jobs differed, and it differed also for different grades in the same job. This difference in mortality may reflect the difference in the nature of job contents, and further studies are warranted to elucidate which job characteristics are responsible.