Nociception monitoring devices are designed to estimate nociception during general anaesthesia. We evaluated the predictive accuracy of heart rate and three nociception indices to predict postoperative pain before emergence from general anaesthesia. In patients undergoing trauma or orthopaedic surgery, HR, Surgical Pleth Index® (SPI), Pupillary Pain Index® (PPI), and Nociception Level® (NOL) were simultaneously recorded for 5 min after the end of surgery but before return of consciousness. After admission to the recovery room, pain scores were assessed regularly for 2 h. HR, SPI, PPI, and NOL were analysed for their predictive accuracy of postoperative pain and opioid consumption with assessment of area under the receiver operating characteristic (AUC) curves, Spearman rank-correlation coefficient, and regression modelling. Data for 60 subjects were analysed. The AUC (95% confidence interval 95% CI) of the predictive accuracy for moderate-to-severe postoperative pain differed between nociception indices (HR=0.46 0.29–0.64, P=0.671; SPI=0.46 0.31–0.61, P=0.621; PPI=0.52 0.36–0.68, P=0.770; NOL=0.66 0.51–0.81, P=0.038). In a multivariable logistic regression model, a higher predictive accuracy was found for a multivariable predictor combining NOL values with ASA physical status and information about use of regional anaesthesia (AUC=0.83 0.72–0.94, P<0.001). Heart rate, Surgical Pleth Index, Pupillary Pain Index, and Nociception Level measured before emergence from general anaesthesia do not yet have sufficient diagnostic accuracy for prediction of postoperative pain. NCT05063227.