Objective To evaluate the reliability, validity and feasibility of the Chinese version of SF-36 health survey questionaire in long-term survivors of nasopharyngeal carcinoma (NPC). Methods A total of 85 long-term NPC survivors completed the Chinese version of SF-36 by either telephone or mail survey. Correlation analysis, reliability analysis and factor analysis were performed to evaluate the reliability and validity of the scale. Results The Chinese version of SF-36 was easy to complete. The split-half reliability was 0. 92 and the Cronbach's α coefficient among domains were all above 0. 70, which showed good reliability and discrimination capacity among domains. All the correlation coefficients between each item and its domain achieved or approached 0. 5, which were greater than those between the item and other domains. These results demonstrated that the Chinese version of SF-36 had good content validity and discriminatory validity. Six principal components were extracted from the scale, which could basically represent eight domains. The cumulative variance was 71.4%. Two common factors were extracted from the eight domains, which accounted for 73.3% of the variance. The Chinese version of SF-36 was able to detect differences in the quality of life between NPC patients and healthy populations. Conclusions The Chinese version of SF-36 has good feasibility, reliability and validity in evaluating the quality of life in long-term NPC survivors.