Sichuan Province plays a crucial role in China’s efforts to achieve carbon neutrality, making it essential to accurately assess the spatial and temporal distribution of carbon dioxide (CO2) concentrations in the region. This study develops a ResNet-LSTM model to address the spatiotemporal fusion of multisource satellite data, specifically integrating CO2 dry air column-averaged mole fraction (XCO2) data from GOSAT, OCO-2, and OCO-3 satellites. By reconstructing the daily spatiotemporal distribution of XCO2 at a 1 km resolution for Sichuan Province from 2015 to 2022, the model fills gaps in satellite observations caused by meteorological conditions and other factors. The results demonstrate significant improvements in accuracy, with the ResNet-LSTM model achieving an R² value of 0.97, outperforming traditional models like XGBoost and Random Forest. The high-resolution XCO2 data provides a robust foundation for validating local emission inventories and supports the formulation of scientifically sound carbon reduction strategies. This study contributes to regional and national carbon neutrality efforts by offering valuable insights into carbon emission dynamics and promoting sustainable low-carbon development.