Objectives : To investigate possible seasonal patterns of pre-term birth in Korea. Methods : A total number of 2,669,357 single live births reported to the National Statistical Office from 1995 to 1998 were analyzed. Composite monthly cohorts of ongoing pregnancies were constructed for each month of the year and the probability of pre-term birth was estimated. Results : Increases in the probability of a pre-term birth occurred during winter for the birth of first child and during summer for the birth of second or later child. This seasonal variation was similar among groups divided by sex, residency, age of mother, and education of mother. Conclusions : These findings suggests that some environmental factors related to season may partially explain the incidence of premature births.