In response to the issues such as sensor noise interference and gyroscope drift during the attitude measurement of low Earth orbit (LEO) satellites，a satellite attitude measurement method based on the Madgwick-extended Kalman filter (EKF) fusion algorithm is proposed.This method uses the data of multiple sensors，e.g.，gyroscopes，accelerometers，and magnetometers，for fusion，and leverages the advantages of both the Madgwick algorithm and the EKF algorithm to achieve attitude measurement.Initially，the Madgwick algorithm is used to calculate the initial attitude with the data measured by multiple sensors.Subsequently，based on the initial attitude and the measured data，the EKF algorithm is used for data fusion and noise filtering so as to obtain the final accurate attitude estimation.The experimental results indicate that compared with the Madgwick algorithm，the fusion algorithm improves the measurement accuracy by 65.8%，and exhibits high robustness.This method provides an effective solution for the attitude measurement of LEO satellites.