An optical phase tracking technique for an extrinsic Fabry-Perot interferometer (EFPI) is proposed in order to overcome interferometric non-linearity. The basic idea is utilizing strain rate information, which cannot be easily obtained from the EFPI sensor itself. The proposed phase tracking system consists of a patch-type EFPI sensor and a simple on-line phase tracking logic. The patch-type EFPI sensor comprises an EFPI and a piezoelectric patch. An EFPI sensor itself has non-linear behavior due to the interferometric characteristics, and a piezoelectric material has hysteresis. However, the composed patch-type EFPI sensor system overcomes the problems that can arise when they are used individually. The proposed system can extract vibration information from severely distorted EFPI sensor signals. The dynamic characteristics of the proposed phase tracking system were investigated, and then the patch-type EFPI sensor system was applied to the active suppression of flutter, dynamic aeroelastic instability, of a swept-back composite plate structure. The real time neural predictive control algorithm effectively reduces the amplitude of the flutter mode, and 6.5% flutter speed enhancement for the aeroelastic system was obtained by integrating smart materials into advanced structures.