Applicability of the cerebellar model articulation controller (CMAC) to structural control is explored, and a training algorithm for the suppression of structural vibration is proposed. The training epochs and controlled responses by both the CMAC and the well-known multilayer neural network are compared in the numerical example. The dynamics of the hydraulic actuator and the time delay of the controller are considered in the simulation. It is concluded that the CMAC is a promising candidate for a structural controller.