Patients with an acute unilateral peripheral vestibular deficit (aUPVD), presumed to be caused by vestibular neuritis, show asymmetrical vestibular ocular reflexes (VORs) that improve over time. Questions arise regarding how much of the VOR improvement is due to peripheral recovery or central compensation, and whether differences in peripheral recovery influence balance control outcomes. Thirty patients were examined at aUPVD onset and 3, 6, and 13 weeks later with four different VOR tests: caloric tests; rotating (ROT) chair tests performed in yaw with angular accelerations of 5 and 20 degrees/s; and video head impulse tests (vHIT) in the yaw plane. ROT and vHIT responses and balance control of 11 patients who had a caloric canal paresis (CP) more than 90% at aUPVD onset and no CP recovery (no-CPR) at 13 weeks in caloric tests were compared with those of 19 patients with CP recovery (CPR) to less than 30%, on average. Balance control was measured with a gyroscope system (SwayStar) recording trunk sway during stance and gait tasks. ROT and vHIT asymmetries of no-CPR and CPR patients reduced over time. The reduction was less at 13 weeks (36.2% vs. 83.5% on average) for the no-CPR patients. The no-CPR group asymmetries at 13 weeks were greater than those of CPR patients who had normal asymmetries. The greater asymmetries were caused by weaker deficit side responses which remained deficient in no-CPR patients at 13 weeks. Contra-deficit side vHIT and ROT responses remained normal. For all balance tests, sway was slightly greater for no-CPR compared with CPR patients at aUPVD onset and 3 weeks later. At 13 weeks, only sway during walking eyes closed was greater for the no-CPR group. A combination of 5 degrees/s ROT and balance tests could predict at onset (90% accuracy) which patients would have no-CPR at 13 weeks. These results indicate that for ROT and vHIT tests, central compensation is observed in CPR and no-CPR patients. It acts primarily by increasing deficit side responses. Central compensation provides approximately 60% of the VOR improvement for CPR patients. The rest of the improvement is due to peripheral recovery which appears necessary to reduce VOR asymmetry to normal at 13 weeks on average. Balance control improvement is more rapid than that of the VOR and marginally affected by the lack of peripheral recovery. Both VOR and balance control measures at onset provide indicators of future peripheral recovery. For these reasons VOR and balance control needs to be tested at aUPVD onset and at 13 weeks.