Objective: The diet of pregnant women is an important factor in the development of the fetus. In our study, we wanted to determine the diet of women who gave birth to healthy children at term (AGA), preterm (PTB) and small for gestational age neonates (SGA). Based on the analysis of dietary components we wanted predict the likelihood of giving birth AGA, PTB and SGA. Methods: The content of components in the women’s diets were estimated based on the dietary questionnaire. The large number of variables analyzed in the diet was reduced using factor analysis. Next, the prediction of prematurity and SGA based on previously selected factors was analyzed. For this purpose, two independent methods were used: discriminant function analysis and ROC analysis. Results: Factor analysis resulted in nine factors containing at least one variable of the factor load being greater than 0.7. Analysis of variance only showed differences between the AGA and preterm groups. The study of discriminant function showed that three factors significantly affect the discriminative power to classify cases into AGA and preterm groups. ROC analysis confirmed diagnostic usefulness factor 1 (fatty acids) in classifying cases into AGA and preterm groups. Conclusion: Based on the analysis of dietary components of women one can predict the likelihood of giving birth to a healthy child at term and prematurely. For AGA the predicting factor is a higher content of short and medium chain fatty acids in a woman’s diet.