This study sought to determine the effective concentration for 50% of the attempts to secure laryngeal mask insertion (predicted EC50LMA) of propofol using a target-controlled infusion (Diprifusor™) and investigated whether fentanyl influenced these required concentrations, respiratory rate (RR) and bispectral index (BIS). Sixty-four elective unpremedicated patients were randomly assigned to four groups (n = 16 for each group) and given saline (control) or fentanyl 0.5, 1 or 2 μg kg−1. Propofol target concentration was determined by a modification of Dixon's up-and-down method. Laryngeal mask airway insertion was attempted without neuromuscular blocking drugs after equilibration had been established for >10 min. Movement was defined as presence of bucking or gross purposeful muscular movement within 1 min after insertion. EC50LMA values were obtained by calculating the mean of 16 patients in each group. Predicted EC50LMA of the control, fentanyl 0.5, 1 and 2 μg kg−1 groups were 3.25 (0.20), 2.06 (0.55), 1.69 (0.38) and 1.50 (0.54) μg ml−1 respectively; those of all fentanyl groups were significantly lower than that of control. RR was decreased in relation to the fentanyl dose up to 1 μg kg−1. BIS values after fentanyl 1 and 2 μg kg−1 were significantly greater than in the control and 0.5 μg kg−1 groups. A fentanyl dose of 0.5 μg kg−1 is sufficient to decrease predicted EC50LMA with minimum respiratory depression and without a high BIS value.