This paper presents an application of a multi-input/multi-output identification technique based on system-generated orthonormal basis functions to a flexible structure. A priori information about the poles of the system, part of which corresponds to the natural frequencies of the structure, is used to generate the orthonormal basis functions. A multivariable model is identified for the experimental flexible structure by using these orthonormal basis functions. It is shown that including a priori knowledge of the system dynamics via the use of orthonormal basis functions into the identification process has the advantage of reducing the number of parameters to be estimated. The multivariable model is used to design an H., controller for the experimental structure to suppress vibrations. The controller is implemented on the structure and very good agreement is obtained between the simulations and the experimental results. (C) 2002 Elsevier Science Ltd. All rights reserved.