A method of measuring dynamic strain signals before and after damage of smart material four-side simple supported plate was developed in this paper.The fiber Bragg grating array and couple neural network were used as senor and signal extraction tool

respectively.The structural damage recognition of the four-side simple supported plate was carried out with coupled neural network method.The results show that the fiber Bragg grating array can sensitively detect the dynamic strain signals.The signal treatment method based on the wavelet packet analysis can obtain a lot of useful characteristic signals with high precision.The couple neural network can get accurate information on location and size of structural damage.The precision of structural damage recognition based on the couple neural network gradually rises with damage size increasing.These results have importantly potential applications in structural health monitoring of large-scale projects.