This study investigates the spatiotemporal characteristics of extreme sea level (ESL) frequency along the China coast using a high-resolution sea-level dataset derived from simulations driven by the Coupled Model Intercomparison Project Phase 6 High Resolution Model Intercomparison Project (CMIP6 HighResMIP) forcing. The exceedance of total sea level above mean higher high water (MHHW) is attributed to five components: sea-level rise (SLR), interannual-to-decadal variability (ID), seasonal cycle (SC), tidal anomaly (TA), and non-tidal residual (NTR). Results show that TA makes the largest median contribution at most stations within the present attribution framework, especially along the South China Sea coast, where historical ESL frequency is strongly modulated by the 18.6-year tidal nodal cycle. ESL seasonality also shows clear regional contrasts. Along the Bohai and Yellow Sea coasts, frequent winter–spring NTR extremes do not lead to a winter ESL maximum because lower seasonal background sea level limits threshold exceedance. Along the Guangdong and eastern Hainan coasts, because SC varies weakly across seasons, ESL occurrence is more strongly controlled by the combined effects of TA and NTR, producing a summer peak and a secondary winter peak. By 2050, 100-year ESLs are projected to increase along most of the China coast, with the largest rises along the Bohai and Yellow Sea coasts, while SLR becomes the dominant driver of future ESL intensification. These results highlight the compound and regionally varying nature of ESL drivers and the growing importance of sea-level rise in future coastal flood risk.