This paper is concerned with optimal experiment design for on-orbit identification of modal frequency and damping parameters in large flexible space structures. The main result is a separation principle for D-optimal design which states that under certain conditions the sensor placement problem is decoupled from the input design problem. This decoupling effect significantly simplifies the overall optimal experiment design determination for large multi-input multi-output structural systems with many unknown modal parameters. The error from using the uncoupled design is shown to be OMICRON ( delta **2) where delta is a measure of the inherent damping of the structure. A numerical example is given demonstrating the usefulness of the simplified criteria in determining optimal designs for on-orbit space station identification experiments.