The British Thoracic Society guidelines for diagnostic flexible bronchoscopy (FB) in adults recommend that intravenous sedation should be offered to patients undergoing bronchoscopy. However, it is difficult to determine the adequate depth of sedation for each patient because of inter-individual variability.This prospective, open-label, single-arm study was conducted in patients undergoing routine bronchus examination with FB. All patients underwent FB under local anesthesia and conscious sedation, with initial administration of 0.03 mg/kg midazolam. The sedation level during FB was objectively assessed using the Ramsay sedation score (RSS). Two hours after the procedure, patients completed a questionnaire about its efficacy and adverse effects using a visual analog scale (VAS). Receiver operating characteristic (ROC) curve analyses were performed to determine the optimal RSS that could improve the subjective efficacy indicated by the VAS.This study enrolled 110 consecutive patients between September 2008 and February 2012. The median total amount of midazolam administered was 1.65 mg per patient. In an analysis of ROC curves between RSS and VAS, the area under the ROC curve for an RSS of 4 against the others was 0.66 (95% CI: 0.54 to 0.77, p = 0.014). The area under the ROC curve was not shown to be statistically significant for RSSs other than 4.The optimal depth of conscious sedation during FB for conventional examination was achieved at an RSS of 4. The patients’ subjective evaluations indicated that a deep level of conscious sedation does not seem necessary for FB.