Climate change has intensified atmospheric conditions favorable to ground-level ozone formation, raising concerns about its impact on public health. This study assessed temporal trends in short-term mortality risk associated with ozone exposure in Seoul, South Korea, from 2008 to 2022. We predicted daily maximum 8-hour ozone concentrations at a 1 km² resolution using a machine learning ensemble model (Random Forest and XGBoost, R² =0.872), incorporating meteorological and observed ozone data. A time-stratified case-crossover design was used to estimate the association between ozone exposure (Moving Average 01) and cause-specific mortality during the warm season (April October). Among 328,804 non-accidental deaths, ozone exposure was significantly associated with all-cause mortality (odds ratio [OR]: 1.01 [1.00-1.01]) and cardiovascular mortality (OR: 1.01 [1.00-1.02]). Respiratory mortality showed a positive but statistically non-significant association. The strength of the association increased over time, with higher ORs and stronger significance observed in later years. Also, a non-linear exposure-response relationship was observed, with a steeper increase in mortality risk at high ozone concentrations. The attributable fraction of deaths on days exceeding the World Health Organization guideline (60 ppb) increased from 4.7% in 2008-2012 to 12.1% in 2018-2022. These results suggest that the health burden of short-term ozone exposure has intensified under ongoing climate change. The findings highlight the importance of continuous air quality monitoring, improved exposure assessment, and strengthened public health strategies, particularly to protect vulnerable populations.