This study compared the stability of the cervical spine according to the presence of neck pain and deep neck flexor performance. Thirty subjects with neck pain, and thirty subjects without neck pain were recruited for this study. The Cranio-cervical flexion (CCF) test was applied using a pressure biofeedback unit to classify the subjects into four subgroups; no cervical pain and good deep neck flexor performance (NG group), no cervical pain and poor deep neck flexor performance (NP group), cervical pain and good deep neck flexor performance (PG group), and cervical pain and poor deep neck flexor performance (PP group). The head sway angle was measured using a three-dimensional motion analysis system. A 3-kg weight was used for external perturbation with the subject sitting in a chair in the resting and erect head positions with voluntary contraction of the deep neck flexors. A one-way analysis of variance (ANOVA) was performed with a Bonferroni post hoc test. The deep neck flexor performance differed significantly among the four groups (p<.05). The NG group had significantly greater deep neck flexor performance than NP and PP groups. The stability of the cervical spine also differed significantly among the four groups in the resting head position (p<.05). The head sway angle was significantly smaller in NG group as compared with the other groups. The PP group had the greatest head sway angle in the resting head position. However, there was no significant difference in the stability of the cervical spine among the groups in the erect head position with voluntary contraction of deep neck flexors (p=.57). The results of this study suggest that the deep neck flexor performance is important for maintaining the stability of cervical spine from external perturbation.