Abstract: According to the linearization error and the susceptibility to bad data in the traditional extended Kalman filter（ EKF）,the noise will dynamically change over time,and it is difficult to accurately obtain it. For this reason,based on the active distribution network,a method based on robust adaptive H∞ extended Kalman filtering（ RAHEKF） is proposed. Based on the original adaptive EKF,the measurement uncertainty theory is adopted,and the measurement evaluation function is introduced to reduce the influence of bad data. The Taylor expansion of the EKF measurement function is retained to the second-order term to reduce the error caused by linearization and enhance the predictive ability of the algorithm under the sudden change of the system. At the same time,the fading memory time-varying noise is used to simulate the change of the noise,and the robustness of the algorithm to the dynamic change of the noise is enhanced. When the improved IEEE 33-node system responds to bad data,sudden changes in system load,and continuous large-scale changes in distributed power,the experimental results show that RAFEKF has higher accuracy and robustness,a and can well adapt to the flexible operation scenario of active distribution network.