We hypothesized that 1) acute severe hypoxia, but not hyperoxia, at sea level would impair dynamic cerebral autoregulation (CA); 2) impairment in CA at high altitude (HA) would be partly restored with hyperoxia; and 3) hyperoxia at HA and would have more influence on blood pressure (BP) and less influence on middle cerebral artery blood flow velocity (MCAv). In healthy volunteers, BP and MCAv were measured continuously during normoxia and in acute hypoxia (inspired O2 fraction = 0.12 and 0.10, respectively; n = 10) or hyperoxia (inspired O2 fraction, 1.0; n = 12). Dynamic CA was assessed using transferfunction gain, phase, and coherence between mean BP and MCAv. Arterial blood gases were also obtained. In matched volunteers, the same variables were measured during air breathing and hyperoxia at low altitude (LA; 1,400 m) and after 1-2 days after arrival at HA (∼5,400 m, n = 10). In acute hypoxia and hyperoxia, BP was unchanged whereas it was decreased during hyperoxia at HA (-11 ± 4%; P < 0.05 vs. LA). MCAv was unchanged during acute hypoxia and at HA; however, acute hyperoxia caused MCAv to fall to a greater extent than at HA (-12 ± 3 vs. -5 ± 4%, respectively; P < 0.05). Whereas CA was unchanged in hyperoxia, gain in the low-frequency range was reduced during acute hypoxia, indicating improvement in CA. In contrast, HA was associated with elevations in transfer-function gain in the very low- and low-frequency range, indicating CA impairment; hyperoxia lowered these elevations by ∼50% (P < 0.05). Findings indicate that hyperoxia at HA can partially improve CA and lower BP, with little effect on MCAv.