This study was undertaken to estimate the incidence of sudden infant death syndrome (SIDS) for 2 years from 1997 to 1998 in Korea. The information on SIDS was obtained from three independent sources, the Korean Medical Insurance Corporation data, data from different medical facilities (Korean surveillance on SIDS) and mortality data from the Korean National Statistical Office. A log linear model with no interactions among the 3 sources was used with the selection criteria of both the Akaike Information Criterion and the Bayesian Information Criterion. The 95% confidence intervals of the estimated number of SIDS deaths were calculated using the goodness-of-fit-method. The estimated number of deaths who were not found in any of the sources was 535, and the estimated total number of SIDS deaths was 759. The 95% confidence interval for estimated total number of patients ranged from 642 to 1522. The estimated rate of SIDS cases was 0.56 per 1000 live births per annum, which is similar to the incidence in the USA or Japan.