OBJECTIVES: From the viewpoint of the health authorities, health-adjusted life-expectancy (HALE) is considered as a comprehensive indicator of mortality and morbidity for monitoring population health over time. A stable, efficient and accurate approach to generate this indicator is demanded. This study aims to compare two approaches, which applied either Euroqol EQ-5D (EQ-5D-3L) or short-form 36 (SF-36) health questionnaire, to estimate HALE in Taiwan. METHODS: In the 2013 National Health Interview Survey, two health questionnaires and data are available from a nationally representative sample. Because there was no localized value sets for these two questionnaires, we applied the Japanese tariff for the EQ-5D-3L and the US preference-based index for the SF-36 in order to compute the health-adjusted weights. Afterwards, the age- and gender-specific health-adjusted weights were incorporated into a current life table according to Sullivan’s method. Two sets of HALE estimates based on either EQ-5D-3L or SF-36 were compared by using the proportion of HALE to life expectancy (LE). RESULTS: In 2013, males had a HALE at birth of 64.21 years (95% confidence interval: 63.65-64.78) based on EQ-5D-3L; it increased to 65.06 years (95% confidence interval: 64.43-65.70) based on SF-36. For females, the HALE at birth were 68.65 years (95% confidence interval: 67.85-69.46) and 67.32 years (95% confidence interval: 66.62-68.01) based on EQ-5D-3L and SF-36, respectively. When we applied the EQ-5D-3L weights, the proportion of HALE to LE was 83% in both gender groups. However, there was difference in the SF-36 weights between males (84%) and females (81%) based on the SF-36. The differences between males and females increased with age. CONCLUSIONS: In our finding, EQ-5D-3L is less sensitive to health status variations due to potential ceiling effect, compared with SF-36. Therefore, SF-36 or other health questionnaires (EQ-5D-5L) could be considered as the major vehicle for estimating HALE in Taiwan.