Vital sign monitoring in neonates requires adhesive electrodes, which often damage fragile newborn skin. Because impulse radio ultrawideband (IR-UWB) radar has been reported to recognize chest movement without contact in adult humans, IR-UWB may be utilized to measure respiratory rates (RRs) in a noncontact fashion. We investigated the feasibility of radar sensors for respiration monitoring in neonates without any respiratory support and to compare the accuracy and reliability of radar measurements with those of conventional impedance pneumography measurements. In the neonatal intensive care unit, RRs were measured using radar (RRRd) and impedance pneumography (RRIP) simultaneously. The neonatal voluntary movements were measured using the radar sensor and categorized into 3 levels (low [M0], intermediate [M1], and high [M2]). RRRd highly agreed with RRIP (r=0.90; intraclass correlation coefficient [ICC]=0.846 [0.835–0.856]). For the M0 movement, there was good agreement between RRRd and RRIP (ICC=0.893; mean bias -0.15 [LOA -9.6–10.0]). However, the agreement was slightly lower for the M1 (ICC=0.833; mean bias=0.95 [LOA -11.4–13.3]) and M2 (ICC=0.749; mean bias=3.04 [LOA -9.30–15.4]) movements than for the M0 movement. In conclusion, IR-UWB radar can provide accurate and reliable estimates of RR in neonates in a noncontact fashion. The performance of radar measurements could be affected by neonate movement.