This study compared the effects of craniocervical flexion exercise with conservative cervical flexion exercise in workers with chronic neck pain. Subjects were randomly allocated to two groups: control (n=20) and experimental group (n=20), respectively. The conservative exercise for the control group consisted of cervical flexion exercise, whereas the experimental group performed a craniocervical flexion exercise. To compare the two groups, the visual analogue scale (VAS) for a pain, neck disability index (NDI) for a neck disability scale, and cross sectional area (CSA) of longus colli muscle using ultrasound image were assessed on pre-intervention, post-intervention, and 2 weeks follow-up. 2-way repeated ANOVA was used with Bonferroni post-hoc test. (1) There were significant main effects (within and between groups) (p<0.05) and interaction effect (p<0.05) in VAS. Post-hoc test revealed that there were significant differences in all pair-wise comparisons. (2) There were significant main effects (within and between groups) (p<0.05) and interaction effect (p<0.05) in NDI. Post-hoc test revealed that there were significant differences between pre-intervention and post-intervention, pre-intervention and 2 weeks follow-up. (3) There were significant main effects (within and between groups) (p<0.05) and interaction effect (p<0.05) in CSA of longus colli muscle. Post-hoc test revealed that there were significant differences in all pair-wise comparisons. The findings indicate that craniocervical flexion exercise are more effective for improving pain, neck disability, and CSA of longus colli muscle than cervical flexion exercise in workers with chronic neck pain.