BACKGROUND: Exposure to non-optimum ambient temperature has been linked to increased risk of total cardiovascular disease (CVD) mortality; however, the adverse effects on mortality from specific types of CVD remain less understood. OBJECTIVES: To comprehensively investigate the association of ambient temperature with cause-specific CVD mortality, and to estimate and compare the corresponding mortality burden. METHODS: We conducted a time-stratified case-crossover study of 1000,014 CVD deaths in Jiangsu province, China during 2015-2019 using data from the China National Mortality Surveillance System. Residential daily 24-hour average temperature for each subject was extracted from a validated grid data at a spatial resolution of 0.0625° × 0.0625°. We fitted distributed lag non-linear models (DLNM) based on conditional logistic regression to quantitatively investigate the association of ambient temperature with total and cause-specific CVD mortality, which was used to further estimate mortality burden attributable to non-optimum ambient temperatures. RESULTS: With adjustment for relative humidity, we observed reverse J-shaped exposure-response associations of ambient temperature with total and cause-specific CVD mortality, with minimum mortality temperatures ranging from 19.5 °C to 23.0 °C. An estimated 20.3% of the total CVD deaths were attributable to non-optimum temperatures, while the attributable fraction (AF) of mortality from chronic rheumatic heart diseases, hypertensive diseases, ischemic heart diseases (IHD), pulmonary heart disease, stroke, and sequelae of stroke was 22.4%, 23.2%, 23.3%, 20.9%, 17.6% and 21.3%, respectively. For total and cause-specific CVDs, most deaths were attributable to moderate cold temperature. We observed significantly higher mortality burden from total and certain cause-specific CVDs in adults 80 years or older and those who were widowed. CONCLUSION: Exposure to ambient temperature was significantly associated with increased risk of cause-specific CVD mortality. The burden of CVD mortality attributable to non-optimum temperature was substantial especially in older and widowed adults, and significantly varied across specific types of CVD.