Premature infants, born before 37 weeks of gestation can have consequences on neurodevelopment and cognition, even when no anatomical lesions are evident. Resting-state functional neuroimaging of naturally sleeping babies has shown altered connectivity patterns, but there is limited evidence on the developmental trajectories of the functional organization in preterm neonates. By using a large dataset (n=597) from the developing Human Connectome Project, we explored the differences in graph theory properties between at-term and preterm neonates at term-equivalent age, considering the age subgroups proposed by the World Health Organization for premature birth. Leveraging the longitudinal follow-up for some preterm participants, we characterized the developmental trajectories for preterm and at-term neonates. We found significant differences between groups for connectivity strength, clustering coefficient, characteristic path length and global efficiency. Specifically, at term-equivalent ages, higher connectivity, clustering coefficient and efficiency are identified for neonates born at higher postmenstrual ages. Similarly, the characteristic path length showed the inverse pattern. These results were consistent for a variety of connectivity thresholds and both at the global (whole brain) and local level (brain regions). The brain regions with the greatest differences between groups include primary sensory and motor regions and the precuneus which may relate with the risk factors for sensorimotor and behavioral deficits associated with premature birth.Our results also show non-linear developmental trajectories for premature neonates, but decreased integration and segregation even at term-equivalent age. Overall, our results confirm altered functional connectivity, integration and segregation properties of the premature brain despite showing rapid maturation after birth.