BACKGROUND: The elderly population has been the primary target of intervention to prevent heat-related illnesses. According to the literature, the highest risks have been observed among the elderly in the temperature-mortality relationship. However, findings regarding the temperature-morbidity relationship are inconsistent. OBJECTIVES: This study aimed to examine the association of temperature with ambulance dispatches due to acute illnesses, stratified by age group. Specifically, we explored the optimum temperature, at which the relative health risks were found to be the lowest, and quantified the health risk associated with higher temperatures among different age groups. METHODS: We used the data for ambulance dispatches in Fukuoka, Japan, during May and September from 2005 to 2012. The data were grouped according to age in 20-year increments. We explored the pattern of the association of ambulance dispatches with temperature using a smoothing spline curve to identify the optimum temperature for each age group. Then, we applied a distributed lag nonlinear model to estimate the risks of the 85th-95th percentile temperature relative to the overall optimum temperature, for each age group. RESULTS: The relative risk of ambulance dispatches at the 85th and 95th percentile temperature for all ages was 1.08 [95% confidence interval (CI): 1.05, 1.12] and 1.12 (95% CI: 1.08, 1.16), respectively. In comparison, among age groups, the optimum temperature was observed as 25.0°C, 23.2°C, and 25.3°C for those aged 0-19, 60-79, and ≥80, respectively. The optimum temperature could not be determined for those aged 20-39 and 40-59. The relative risks of high temperature tended to be higher for those aged 20-39 and 40-59 than those for other age groups. CONCLUSIONS: We did not find any definite difference in the effect of high temperature on ambulance dispatches for different age groups. However, more measures should be taken for younger and middle-aged people to avoid heat-related illnesses.