To determine the effects of endogenous and exogenous strogen on serum lipid levels, twenty nonsmoking healthy Korean women were participated in this experiment for 12 weeks. They were assigned to three groups : (1) eight women aged 22 to 30(yr) for the premenopausal(Pre) group, (2) eight, aged 49 to 60(yr) for the postmenoparusal(Pst) group, (3) four, aged 23 to 30(yr) for the oral contraceptive(OC) group which used triphasic OC formulation. Fasting blood samples representing every phase of the hormonal levels were obtained from the subjects of the Pre and the OC group. From the subjects of the Pst group, fasting blood samples were obtained once per three weeks for 12 weeks. All the serum data were adjusted for dietary effects, exercise, personality type and body mass index(BMI) by using analysis of covariation(ANCOVA). Serum lipid levels of the three groups were significantly different. While serum levels of triglycerides(TG)(p<0.0001), low density lipoprotein-chloesterol(LDL-C)/high density liporotein-cholesterol(HDL-C) ratio (LDC-C/HDL-C)(P<0.01) and total cholesterol (TC)/HDL-C ratio (TC/HDL-C)(P<0.001) were significatnly high in the Pst group, serum HDL-C(P0.001) level was significantly high in the Pre group. The OC group showed significantly low serum TC(P<0.0001) and LDL-C(P<0.0001) levels. There was no signidicant difference in the fluictuation of serum lipid levels during the menstrual cycle of the Pre group. However, in the OC group, serum TG level was significantly increased at phase 2(P<0.05) where exogenous estrogen administration was highest. Even though other serum lipid levels of the OC group were not significantly fluctuated according to the exogenous estrogen administration, there was a trend of increased levels of serum TC, LDL-C, LDL-C/HDL-C and TC/HDL-C and decreased level of HDL-C during the menstruation period. Also, serum TC level was high(P<0.005) and serum TG level was low (P<0.005) at the baseline of the OC group compared with the periods of OC administration. When screening and counseling the female population at risk for coronary heart disease(CHD), the result of this study suggest that in may be desirable to divide the population into several groups according to their personal physiological characteristics, such as age, OC administration, menstrual cycle and menopause, as well as general risk factors for CHD.