Aim: To compare the safety and efficacy of dexmedetomidine and haloperidol in patients with delirium and acute respiratory failure in non-invasive mask lung ventilation.Materials and methods. We carried out a retrospective analysis of data on patients with somatogenic delirium and acute respiratory failure (57 men, 4 women) experienced noninvasive ventilation (NIV) in the intensive care unit of the Botkin State Clinical Hospital in 2017–2018. Depending on the type of sedation the patients were divided into two groups: those on dexmedetomidine (n=31) and those receiving haloperidol (n=30). Dexmedetomidine was administered as a continuous infusion at a rate of 0.2–1.4 µg/kg/h while controlling the level of consciousness; haloperidol was administered by intravenous bolus injections until a sufficient level of sedation was reached in a dose of 2.5 mg 2–3 times a day.Results. The efficiency of sedation to achieve the required level of cooperation and possibility of NIV was 87.1% (27 patients) and 66.6% (20 patients) in dexmedetomidine and haloperidol groups. When estimating sedation in patients of both groups according to the RASS scale the scores did not differ significantly and were equal on the average to 1.7±0.3 (eye contact to voice). In haloperidol group in 10 out of 30 (33.3%) patients a sufficient level of sedation was not achieved, which required immediate tracheal intubation and invasive lung ventilation. Mortality in this group was 20% (6 patients), while in dexmedetomidine group it was 6.4% (2 patients).Conclusion. The use of dexmedetomidine, despite greater variability of hemodynamic parameters, allows to perform NIV with sufficient cooperation with the patient, reduces the frequency of tracheal intubation, risk of complications and mortality.