Objectives: Due to global warming resulting from climate change, there has been increasing interest in the relationship between temperature and mortality. These temperature-related deaths depend on diverse conditions related to a given place and person, as well as on time. This study examined changes in the impact of high temperatures on death in summer, using the effect and burden of elevated temperatures on deaths in Seoul and Daegu. Methods: A Poisson regression model was used to estimate short-term temperature effects on mortality. Temperature-related risks were divided into three time periods of equal length (1996-2000, 2001-2005, and 2006-2010). In addition, in order to compare the impact of high temperatures on deaths, this study calculated the proportion of attributable deaths to population, which simultaneously considers the threshold and the slope above the threshold. Results: The effect and burden of high temperatures on deaths is high in Daegu. However, the impact (i.e. the effect and burden) of elevated summer temperatures on deaths has declined over the past 15 years. Sensitivity analyses using alternative thresholds show the robustness of these findings. Conclusion: This study suggests that the attributable burden of high temperatures on deaths to be more plausible than relative risk or threshold for comparing the health impact of high temperatures across populations. Moreover, these results contain important implications for the development or the adjustment of present and future strategies and policies for controlling the temperature-related health burden on populations.