Life table of Korean population for years 1983 1985 1987 and 1989 were constructed byt the National Bureau of Statics. The ago specific death rates were calculated froom the death registral ion for numerators and the estimated population by age and sex for denominators. In the course of constructing life tables, we have maole some adjustments for deficiencies in regist rat ion olata as follows. First, the non-registered portion oof infant deaths especially for neo-natal deaths was estirnateol and added too the original data. The main reason is that deaths occorring in the neo-natal period and prior to the registrat ion of birth leave little incentive for the registration of either the birth or the death. Second. t he do~hayed p(ortioon of deaths registering after one year of occurrence was estimated and added too the original data. Third t the ptortioon haying in, occuracies in ,~oge reporting was also estimated. Fourth the moving average methood was finally employed in an effort too remove the random error. The major fin(hings are as foolloows. 1. the average life expectancy at birth in 1989 is calculateol as 70.8 years in 1989, 2. a gap netween the male and female life expectancies is widened to more than 8 years toorm 1.8 years in 1906 10. It means that the female life expectancy has increased substantially, 3. the death rates of the middle - aged men starting age 40 are found to he relatively higher than those of females and younger age groups. This peculiar pattern was also found with the comparison of those of other countries.