Background: The efficacy and safety of dexmedetomidine and olanzapine for delirium control in critically ill elderly patients without ventilation or surgery are not known. Methods: The efficacy and safety of dexmedetomidine and olanzapine for controlling delirium were evaluated in a retrospective cohort of critically illness by assessing the sedation level, drug dose/duration, combination rate with other sedatives, adverse effects, intubation rate and prognosis. Result: The maximum (1.61 +/- 1.56 vs. 2.70 +/- 1.01, p < 0.001), average (-0.57 +/- 0.88 vs. 0.88 +/- 0.73, p < 0.001), and minimum (-1.67 +/- 1.04 vs. -1.37 +/- 1.01, p = 0.014) RASS scores of 263 patients were lower after treating with dexmedetomidine than treating with olanzapine. Drug use duration (4.83 +/- 2.67 days vs. 5.87 +/- 3.14 days, p = 0.005) and sedative combination rates (13.56% vs. 40.00%, p = 0.003) were lower when treating with dexmedetomidine than that with olanzapine. A comparison of adverse effects between dexmedetomidine and olanzapine revealed respiratory depression (16.95% vs. 2.84%, p < 0.001), hypoxia (13.56% vs. 2.76%, p < 0.001) and hypotension (11.02% vs. 3.45%, p = 0.007). Intubation rates (22.88% vs. 12.41%, p = 0.023) and the length of hospital stay (9.30 +/- 4.90 days vs. 8.83 +/- 3.34 days, p < 0.001) were higher in patients treated with dexmedetomidine than that with olanzapine. Mortality rates, cognitive prognosis, and delirium recurrence rates were similar between groups. Age, severe cardiopulmonary disease, APACHE II scores, dexmedetomidine dose, minimum RASS score and sedative combination were significantly (p < 0.05) associated with the adverse effects of dexmedetomidine. Respiratory depression, hypoxia and hypotension in the olanzapine group all occurred during combination with benzodiazepines. Conclusions: Dexmedetomidine achieved more satisfactory sedative effects on delirium control, but olanzapine was safer.