The aim of this study is to identify the effects of hand positions (head and jaw) on the video head-impulse test (vHIT). Eighty-six healthy volunteers and sixty-seven patients with unilateral vestibular neuritis (UVN) were recruited for this study. Different hand positions (head and jaw) were used in the vHIT of horizontal semicircular canals in healthy volunteers and UVN patients. All the obtained horizontal vHIT gains were analyzed. It was observed that when horizontal vHIT was performed with the head hand position, the number of head impulses that produced overhigh vestibulo-ocular reflex (VOR) gains was more than that with the jaw hand position ( < 0.01), irrespective of whether the test was performed in healthy volunteers or UVN patients. The gains obtained were lower when the jaw hand position was used than that obtained when the head hand position was used ( < 0.05). However, no significant difference existed in the mean head velocity between the two hand positions ( > 0.05). Using the head hand position has greater a chance to elicit in UVN patients normal horizontal vHIT gains with refixation saccades than using the jaw hand position ( = 0.04). The jaw hand position can increase the accuracy of vHIT in determining the lesion side.