Objectives: To estimate the association of work-related musculoskeletal neck symptom and occupational risk factors, and to identify the dose-response relationships between work-related musculoskeletal neck symptom and physical condition. Methods: Workers from four metal industry companies were enrolled in the study. General characteristics, psyhcosocial and physical risk factors and work-related musculoskeletal neck symptom were investigated by questionnaire in the baseline study. At the follow-up study 2 years later, the neck symptom was surveyed. Workers who did not complain neck symptom at baseline and complain at follow-up, and workers who did complain neck symptom at baseline and follow-up were defined as case. The adjusted odds ratios (ORs) of physical risk factors were estimated for these subjects. Additionally, subjects who did not change their work at follow-up were analysed. Results: Among the cohort of 251, 199 (79.3%) were used in the analysis, of which 71 (35.7%) were classified as cases. The adjusted ORs and the ORs of the subjects who did not change their work were increased with exposure to the following kinds of work: ‘repetitive’, ‘lift’, ‘push/pull’, ‘neck/shoulder (overhead/bend)’, ‘hand/wrist (bend/pinch)’, ‘trunk (twist/bend)’, ‘pressure point’, and ‘incentive work’. Finally, with increasing exposure to ‘repetitive’, ‘lift’, ‘push/pull’, ‘awkward postures’, ‘pressure point’, and ‘incentive work’, the risks of work-related musculoskeletal neck symptom were increased. Conclusions: Work-related musculoskeletal neck symptom was increased with increasing exposure time of certain physical risk factors, and a dose-response relationship was observed.