Purpose This study aims to investigate whether the usage of selective COX-2 inhibitors would reduce the risk of recurrence in male post-operative head and neck (H& N) cancer patients. Methods This study population is based on the National Health Insurance Research Database (NHIRD) in Taiwan. A number of 31,020 newly diagnosis male H& N cancer patients was retrieved by ICD-9-CM (140–141, 143–146, 148–149, 161) during 1998–2007. And, 22,521 patients who received surgery within 4months after first diagnosis were further selected. The usages of selective COX-2 inhibitors within the year after diagnosis were computed and transformed into define daily dose (DDD) and medication procession ratio (MPR). Cancer recurrence was defined as a new treatment of surgery, radiotherapy (RT) or chemotherapy (CT) identified from database as the first appearance of treatment after a 180-day free of surgery, RT or CT. The proportional hazards model using cancer recurrence as the endpoint was used to estimate the effects from selective COX-2 inhibitors. Analysis covariates also included age, gender, living area, occupation and comorbidities in the year before cancer diagnosis, and other treatments of CT, RT or in combination, before or after surgery. Results Of the 11,357 patients with late recurrence, defined as those had recurrence 2 or more years after primary treatment, the proportional hazards model analyses indicated patients with the history of selective COX-2 inhibitors in lower MPR (less than 80%) had significantly lower risk of cancer recurrence (HR=0.80, 95%CI=0.69–0.92, p=0.002) than those without selective COX-2 inhibitors usage. The risk of recurrence was not significant affected in the higher MPR users (HR=0.86, 95%CI=0.35–2.11, p=0.740). Conclusions Of male H& N cancer patients, the usage of selective COX-2 inhibitors (MPR less than 80%) during the year of cancer diagnosis is associated with 20% risk reduction in late (⩾2years) cancer recurrence after primary treatment.