Background Automatic lancets have been reported to be superior to manual lancets in terms of pain and treatment time. However, no studies have yet been published comparing automatic lancet and needle puncture heel‐prick blood sampling. The objective of this study was to compare the pain response and efficiency between the automatic lancet and needle at the time of heel blood sampling. The design was a randomized controlled trial. The inclusion criteria for the participants were a birthweight of ≧1,500 g and a gestational age of ≧30 weeks. Methods The study examined a total of 105 neonates who were randomized into an automatic lancet group (n = 53) and a needle group (n = 52). The parameters measured included blood collection time, number of calf squeezes, duration of audible crying, and the Neonatal Infant Pain Scale (NIPS) score. The main outcome measure was audible crying duration. Results The duration of audible crying was significantly shorter in the automatic lancet group when compared to the needle group (median 3 s, interquartile range (IQR) 0–33 s vs median 39 s, IQR 5–91.5 s, P = 0.0023). The NIPS score at the time of puncture was significantly lower in the automatic lancet group than in the needle group (median 1, IQR 0–5 vs median 5, IQR 3–6, P = 0.0060). There was no significant difference in the blood collection time and the number of calf squeezes between the two groups. The automatic lancet was found to be less painful than the needle puncture in neonatal heel‐prick blood sampling with no significant difference in blood sampling time. Conclusion The automatic lancet was found to be less painful than the needle puncture in neonatal heel‐prick blood sampling with no significant difference in blood sampling time.