Background: The microbial population living in our intestines plays a key role in several metabolic, nutritional, physiological, and immunological processes. It is known that infant gut microbiota composition has both horizontal transmission delivery and environmental conditions and a vertical one, from mother to child, related to how the infant is fed (breastfed or infant formula). Detailed information on the composition of meconium and feces from newborns may help predict the most prevalent and hazardous conditions affecting pregnancies, mothers, and babies, including pre-term birth, preeclampsia, and gestational diabetes mellitus for example. Methods: This work aims to demonstrate the feasibility of the whole High-Resolution proton Nuclear Magnetic Resonance (1H HR NMR) procedure in metabolomic analysis in preterm newborns. Thus, multiple samples of meconium and stool were collected from 3 pairs of premature twins and their metabolite profiles were acquired and exploited by combining the NMR technique with univariate and multivariate analysis. Results: The analysis showed that an impact on the metabolite profile was visible concerning both the sex of the newborns and the couplet of origin. Most of the variation between twin couplets was seen with butyric acid concentration in meconium/stool samples. Conclusions: Despite the low number of samples, the described NMR procedure showed to be a suitable approach to evaluate the similarities of the molecular profiles of different samples, offering a non-invasive and informative approach to understanding the metabolic and nutritional status of preterm infants. Future metabolomic analysis should be supported by microbiome analysis, such multiomic approach will provide more complex view on development of preterm newborns.