Amid the accelerated growth in global atmospheric methane (CH₄) concentrations after 2019, identifying key emitting regions, quantifying their contributions, and elucidating the underlying drivers have become pressing needs. However, limited monitoring capacity and complex inversion systems have constrained the timely and accurate assessments of regional CH₄ emissions. Here, we construct a regional atmospheric inversion framework using the Local Ensemble Transform Kalman Filter (LETKF), constrained by satellite CH₄ observations. Applied to East Asia at 0.5° × 0.625° resolution, this system produces weekly CH₄ flux estimates for China during 2019–2024. We show that China’s CH₄ emissions increased from 61.1 (56.2–66.7) Tg in 2019 to 66.8 (61.5–73.0) Tg in 2024. The livestock sector contributed nearly half of the growth, while rising waste and oil-gas emissions and northward expansion of rice cultivation shifted China’s emissions growth to previously low-emitting Northwest and Northeast regions. Our framework demonstrates the feasibility of near-real-time, regional-scale emissions monitoring, offering a transferable tool for other high-emitting countries.