Background: Virtual reality (VR) applied to patients with neck pain is a promising intervention to produce positive effects when used alone or combined with exercise. Therefore, the objective of this manuscript is to compare the effects of VR versus exercise treatment on pain intensity, conditioned pain modulation (CPM), temporal summation (TS) and functional and somatosensory outcomes in patients with non-specific chronic neck pain (NS-CNP). Methods: A single-blinded, randomized clinical trial was carried out. A total sample of 44 patients with NS-CNP was randomized into a VR treatment group or neck exercises group. The intervention consisted of two treatment sessions per week, for four weeks and eight sessions. Four measurement moments (at baseline, immediately, 1 month, and 3 months after intervention) were considered. Pain intensity, CPM, TS, functional and somatosensory outcomes were measured. Results: Statistically significant differences were revealed for time factor (F = 16.40, p < 0.01, ηp2 = 0.28) and group*time interaction for kinesiophobia (F = 3.89, p = 0.01, ηp2 = 0.08) showing post-hoc differences in favor of the VR group at 3 months (p < 0.05, d = 0.65). Significant effects were shown for time factor (p < 0.05) but not for the group*time interaction (p > 0.05) for pain intensity, rotation range of motion (ROM), Neck Disability Index, pain catastrophizing, fear-avoidance beliefs, left side pressure pain threshold (PPT) and anxiety. Statistically significant differences were not found for time factor (p > 0.05) and neither in group*time interaction (p > 0.05) for CPM, TS, right side PPT, flexo-extension and lateral-flexion ROM. Conclusions: Kinesiophobia was the only outcome that showed differences between VR and exercise at 3 months. Nevertheless, pain intensity, CPM, TS, ROM, neck disability, pain catastrophizing, fear-avoidance beliefs, PPT and anxiety did not show differences between both interventions.