Increased mortality, linked to events of extreme high temperatures, is recognized as one critical challenge to the public health sector. Therefore, this ecological study was conducted to assess whether this association is also significant in Taiwan and the characteristics of the relationship. Daily mean heat indices, from 1994 through 2008, were used as the predictor for the risk of increased mortality in populations from 6 major Taiwanese cities. Daily mortality data from 1994 through 2008 were retrieved from the Taiwan Death Registry, Department of Health, Taiwan, and meteorological data were acquired from the Central Weather Bureau. Poisson regression analyses using generalized linear models were applied to estimate the temperature-mortality relationship. Daily mean heat indices were calculated and used as the temperature metric. Overall, increased risk ratios in mortality were associated with increased daily mean heat indices. Significantly increased risk ratios of daily mortality were evident when daily mean heat indices were at and above the 95th percentile, when compared to the lowest percentile, in all cities. These risks tended to increase similarly among those aged 65. years and older; a phenomenon seen in the cities of Keelung, Taipei, Taichung, Tainan, and Kaohsiung, but not Chiayi. Being more vulnerable to heat stress is likely restricted to a short-term effect, as suggested by lag models which showed that there was dominantly an association during the period of 0 to 3. days. In Taiwan, predicting city-specific daily mean heat indices may provide a useful early warning system for increased mortality risk, especially for the elderly. Regional differences in health vulnerabilities should be further examined in relation to the differential social-ecological systems that affect them.