Purpose: To investigate the effect of several ranges of assistance levels on respiratory pattern including peak inspiratory pressure ( PIP), tidal volume (TV), electrical activity of diaphragm (Edi), work of breathing and comfort during neurally adjusted ventilatory assist (NAVA) in preterm infants. Methods: A prospective observational study was performed to preterm infants born before 36 weeks of gestation, received ventilatory support at least 24 hours. Seven increasing and decreasing NAVA levels (from 0.5 to 4.0 with interval of 0.5) were applied each for 10 minutes and the ventilator data were recorded for last 5 minutes. Premature infant pain profile ( PIPP) scores were calculated every 10 minutes. Results: Twelve preterm infants were included. Median (IQR) gestational age and birth weight was 29.4 (26.1-34.0) weeks and 1180 (895–2160) g, respectively. The prior duration of mechanical ventilation was median 10 (IQR 3-19) days. At the study point, their median (IQR) postconceptional age, postnatal age and weight was 33.6 (31.1-34.5) weeks, 9 (2.5-34) days and 1650 (1360-2275) g, respectively. Baseline median (IQR) PIP, TV and Edi was 12.6 (10.5-14.5) cmH2O, 6.2 (4.9-7.5) mL/kg and 12.4 (6.7-15.0) µV at NAVA level 0.5. According to the stepwise raising of NAVA level from 0.5, TV increased after 1.0 rise of NAVA level. Contrary, Edi decreased after 1.5 level rise. During the NAVA level reduction from 4.0, while TV decreased after 0.5 reduction, Edi showed significant increase after 2.5 level reduction. Similarly, whereas efficacy indices ( PIP/Edi and TV/Edi) reflecting unloading of respiratory support increased after small rise of NAVA level (0.5 to 1.0) during assistance raising, they decreased after 2.5 reduction during NAVA level weaning. As an aspect of pain and discomfort, PIPP scores moved reciprocally to the change of NAVA level (p=0.036) and its difference from the baseline was significantly correlated with Edi change (r=0.77, p=0.005). Conclusion: Raising NAVA levels affected in not only increase of TV, also decrease of peak Edi, work of breathing and pain scores in preterm infants. Unlike raising NAVA levels, Edi and work of breathing increased after more elimination of assistant level during NAVA weaning. NAVA could effectively unload the respiratory work and reduce the consequential pain and discomfort in preterm infants.