CO2 is a primary driver of climate change, with significant contributions from agricultural landscapes like the rice paddies of Vietnam. Accurate, high-resolution monitoring of CO2 concentrations in these areas is crucial for climate mitigation, yet ground-based measurements are sparse. This study presents a computational framework to estimate column-averaged CO2 (XCO2) emissions from rice fields by integrating data from NASA’s Orbiting Carbon Observatory-2 (OCO-2) with satellite-derived vegetation indices and land cover information. The approach combines column-averaged CO2 measurements with biophysical and meteorological variables, processed using an optimized Support Vector Regressor (SVR) machine learning model. The model successfully captured the seasonal carbon cycle, which aligns with Vietnam’s rice cultivation calendar. This demonstrates the framework’s potential for monitoring GHG dynamics in response to agricultural activities. This approach offers a scalable, data-driven tool to support sustainable land management and climate mitigation efforts in critical food-producing regions.