To explore the effect of giving sedatives according to the circadian rhythm in prevention of occurrence of delirium and the prognosis of patients undergoing mechanical ventilation in intensive care unit (ICU). A prospective double-blinded randomized controlled trial (RCT) was conducted. The patients admitted to Department of Critical Care Medicine of the Second Hospital of Lanzhou University from July 2014 to February 2015, undergoing invasive mechanical ventilation over 12 hours were enrolled. All the patients were given fentanyl for analgesia, and they were randomly divided into simulated circadian clock group (study group, n = 35) and non-simulated circadian clock group (control group, n = 35). The patients in each group were subdivided into three subgroups according to the kinds of sedative drugs, namely dexmedetomidine group (n = 8), propofol group (n = 14), and dexmedetomidine combined with propofol group (combination group, n = 13). Visual analogue scale (VAS) standard and Richmond agitation-sedation scale (RASS) were used to control the analgesic and to quantify the depth of sedation by titrating the dose of sedative drugs, the simulated circadian clock was set to control the RASS score at 0-1 during the day, and -1 to -2 at night in study group. The RASS score in the control group was set at -1 to -2 day and night. The urine 6-hydroxy acid melatonin (aMT6s) levels at different time points in the first diurnal rhythm (06:00, 12:00, 18:00, 24:00) were determined by enzyme linked immunosorbent assay (ELISA). The incidence of delirium, severe hypotension, severe bradycardia and other adverse reactions, duration of mechanical ventilation and the time of extubation, length of ICU stay, amount of sedative and analgesic drugs used were recorded. The correlation between delirium and other indexes was analyzed by using Spearman correlation analysis. (1) There were no significant differences in gender, age, acute physiology and chronic health evaluation II (APACHEII) score among groups. (2) Urine aMT6s levels did not show circadian rhythm in both groups, aMT6s level at 06:00 in study group showed an increasing tendency as compared with the control group, but the difference was not statistically significant. (3) Compared with the control group, the incidence of delirium was significantly lowered in the study group (14.3% vs. 37.1%, P = 0.029), but no significant differences were found in the incidence of severe hypotension or severe bradycardia (20.0% vs. 25.7%, 11.4% vs. 20.0%, both P > 0.05). In simulated circadian clock group, the incidence of delirium in dexmedetomidine group was significantly lower than that of the propofol group (6.3% vs. 32.1%, P < 0.05). (4) Compared with control group with the same sedative, the duration of mechanical ventilation, extubation time, length of ICU stay were significantly shortened, and the dosage of sedative drugs used was reduced in study group (all P < 0.05). In simulated circadian clock group, the duration of mechanical ventilation in dexmedetomidine group was significantly shorter than that of propofol group and combination group (hours: 75.75±26.78 vs. 102.00±26.31 and 100.31±25.38, both P < 0.05), and the length of ICU stay was significantly shorter than that of propofol group (days: 5.75±1.04 vs. 7.00±1.52, P < 0.05). (5) The occurrence of delirium was positively correlated with duration of mechanical ventilation (r = 0.705), extubation time (r = 0.704), length of ICU stay (r = 0.666, all P = 0.000), and no correlation was found between the occurrence of delirium and aMT6s level at 06:00, 12:00, 18:00, and 24:00 (r = -0.135, r = 0.163, r = 0.269, r = -0.077, all P > 0.05). Administration of sedatives according to simulating circadian time could decrease the duration of mechanical ventilation, extubation time, and the length of ICU stay, decrease the dosage of sedative drugs, and reduce the incidence of delirium. Dexmedetomidine could reduce the incidence of delirium, and improve the prognosis of patients.