Background: The analysis of infant mortality from congenital malformations, which at present is the main group of causes of this mortality in Chile, suggests that it could be decreased with a good knowledge of its conditioning factors. Aim: To study infant mortality differentials from congenital malformations with linked records, in the 1993 to 1995 Chilean birth cohorts. Material and methods: Analysis of mortality differentials in 1993,1994 and 1995 birth cohorts. Multivariate logistic regression of mortality from congenital diseases. Results: Univariate analysis showed that mortality is highest in the Southern regions of the country (VII to XII) and in rural areas. It is also higher in children from older and from very young mothers, it increases along with the birth order of the child and decreases with increasing educational level of the mother. Multiple logistic regression analysis, confirmed the higher mortality in the Southern regions, aged mothers, high birth order of the child and low educational level of the mother. However no significant influence of rurality nor greater mortality in children of very young mothers was found. Conclusions: These results can be attributed to the fact that this type of analysis permits the control with other variables. Although the mortality data showed interesting relationships with the independent variables, a registry of all live births and stillbirths with congenital anomalies, that would provide greater numbers and data on non fatal anomalies, would be desirable to better study their causal factors. (Rev Méd Chile 2001; 129: 405-12)