Aiming at the robustness and stability of the vibration active control of motor-driven seawater pumps， a hybrid structure adaptive vibration active control strategy is proposed based on the Kalman filter （KF） algorithm， which establishes the system state prediction equations， state transfer matrix and measurement matrix， and builds a hybrid structure adaptive vibration active control system model. In order to improve the convergence performance of the algorithm， an online update strategy for the mea surement noise covariance matrix is proposed. Simulation results show that the new control strategy effectively overcomes the strong correlation between the reference signal and the vibration source based on the classical Filtered x Least Mean Square algo rithm （“FxLMS”）， and realizes effective vibration active control under the premise that Gaussian white noise is used as the refer ence signal. The robustness， stability， and control effect of the proposed strategy are all superior to that of the FxLMS algorithm with a variable step size. The results provide theoretical support for engineering practice and have certain potential application value.