Summary This trial aimed to compare the maternal and neonatal effects of remifentanil given by patient‐controlled analgesia (PCA) or continuous infusion for labour analgesia. Patient controlled analgesia was administered using increasing stepwise boluses from 0.1 to 0.4 μg.kg −1 (0.1 μg.kg −1 increment, 2 min lockout, n = 30). Continuous infusion used rates from 0.05 to 0.2 μg.kg −1 .min −1 (0.05 μg.kg −1 .min −1 increment, n = 30). Dose increments were given on request. Women reported lowest pain scores (median ( IQR [range]) of 3 (2–4 [2–5]) for PCA and 4 (3–5.25 [3–7]) for continuous infusion (p = 0.004) at 60 min after the beginning of analgesia. The mean ( SD ) remifentanil umbilical vein/maternal artery ratio in the PCA and infusion groups were 0.74 (0.45) vs 0.70 (0.52), respectively (p = 0.776). The mean ( SD ) umbilical artery/umbilical vein ratios were 0.31 (0.12) vs 0.26 (0.07), respectively (p = 0.088). Maternal and neonatal adverse reactions of remifentanil were similar between the two groups. The total remifentanil consumption (median ( IQR [range]) during PCA administration was lower than continuous infusion, 1.34 (1.22–1.48 [0.89–1.69]) mg vs 1.49 (1.35–1.61 [1.12–1.70] mg; p = 0.011). The results suggest that remifentanil PCA provides better pain relief and similar placental transfer compared with continuous infusion.