Objective Consensus around the ideal chart to classify preterm babies is scant. It is particularly relevant in small for gestational age (SGA) infants due to its clinical and therapeutic implications. The aim of the study was to compare Olsen, Intergrowth-21st, and Fenton growth charts, regarding the classification at birth and incidence of SGA preterm infants. Study Design Retrospective study of 529 preterm infants ≤ 32 weeks of gestational age. Birth weight Z-score was calculated applying the three growth charts and ponderal index (PI) was also estimated. Incidence of SGA (birth weight < 10th percentile) and clinical outcome were compared according to the chart used. Results Incidence of SGA was significantly higher (p < 0.001) with Olsen (101 cases, 19.1%) compared with Intergrowth-21st (75 cases, 14.2%) and Fenton (53 cases, 10%). Differences were also found with PI of SGA preterm infants, as those infants classified by Olsen were mostly symmetric (PI > 10th percentile), while Fenton and Intergrowth-21st identified less symmetric SGA infants. Kappa concordance between Intergrowth-21st and Fenton was 0.805, Intergrowth-21st versus Olsen 0.824, and Fenton versus Olsen 0.641. No differences were observed on neonatal morbidities or mortality. Conclusion Significant differences were detected when classifying very preterm infants at birth according to the growth chart, mainly among symmetric SGA. Concordance between Fenton and Olsen was poor, but Intergrowth-21st showed high concordance with Fenton and Olsen. However, further research is needed to select the ideal chart. Variability in the population selected to create the curves and the accuracy dating the pregnancy are factors that may have explained differences. Key Points