All realistic engineering systems operate in uncertain environments, and in some cases even small levels of noise can severely compromise their performance and stability. In this paper, we propose a robust design approach to improve the control performance of dynamic systems. Our proposed approach first specifies a set of linear constraints describing dynamic performance requirements based on pole placement, and develops a robust linear programming model which is solved to yield a parameter design that achieves the performance requirements under uncertainties. The advantage of our proposed approach is its computational efficiency since it reduces the problem to solving a small sequence of linear programming problems, and the proposed optimisation model can be embedded easily in a local search framework. The effectiveness of our proposed approach is verified with theoretical developments and computational studies on a typical high-performance mechatronic system; the hard disc drive.