The effect of exposure to extreme temperature events (ETEs) on dementia mortality remains largely unknown. We aimed to quantify the association of ETE exposure with dementia mortality. We conducted a population-based, case-crossover study among 57 791 dementia deaths in Jiangsu province, China, during 2015-20. Daily mean temperatures were extracted from a validated grid dataset at each subject's residential address, and grid-specific exposures to heat wave and cold spell were assessed with a combination of their intensity and duration. We applied conditional logistic regression models to investigate cumulative and lag effects for ETE exposures. Exposure to ETE with each of all 24 definitions was associated with an increased odds of dementia mortality, which was higher when exposed to heat wave. Exposure to heat wave (daily mean temperature ≥95th percentile, duration ≥3 days (d); P95_3d) and cold spell (≤5th percentile, duration ≥3 d; P5_3d) was associated with a 75% (95% CI: 61%, 90%) and 30% (19%, 43%) increase in odds of dementia mortality, respectively. Definitions with higher intensity were generally associated with a higher odds of dementia mortality. We estimated that 6.14% of dementia deaths were attributable to exposure to heat wave (P90_2d) and cold spell (P10_2d). No effect modifications were observed by sex or age, except that the association for heat wave was stronger among women. Exposure to both heat wave and cold spell was associated with an increased odds of dementia mortality. Our findings highlight that reducing individual ETE exposures may be helpful in preventing deaths from dementia, especially among women in summer.