Background/Aims: Midazolam is generally used for sedation during gastroscopy, while propofol is usually used for outpatients’ day surgery anesthesia. We compared propofol plus fentanyl with midazolam for sedation during gastroscopy. Methods: One hundred and fifty three patients who had requested sedation during diagnostic gastroscopy were randomly divided into two groups. One group (n=98) received an initial bolus dose of propofol (40 mg) plus fentanyl (50 μg) intravenously, followed by additional doses of propofol (10 mg) at one minute interval (PF group) until achieving sedation. The other group (n=55) received an initial bolus dose of midazolam (3 mg) intravenously, followed by additional doses of midazolam (1 mg) at two minutes interval (M group). Results: Sedation was achieved faster in PF group than in M group. Time to regiain orientation and recover warking was also shorter in PF group than in M group. Gag reflex and procedure difficulty were less in PF group than in M group. Incidences of complications were not different between the two groups. Degree and frequency of oxygen desaturation were higher in PF group than in M group. Conclusions: Propofol plus fentanyl seems to be more acceptable and suitable than midazolam for sedation during gastrointestinal endoscopic examination. (Korean J Gastroenterol 2001;38:98-105)