Abstract: In this paper, a delamination identification strategybased on vibration signatures by using artificial neural networks ispresented Through different nodes with same coordinates, a finite elementmodel for internal delamination is established.The global vibrationidentification factors of damaged and damage free laminates are obtained bynumerical simulation. In the studies, the identification capability for thequantitative prediction of delamination in composites laminates is focused.The results show that the strategy based on the vibration signal measurementand artificial neural networks is efficient to detect delamination defectsand there is a good possiblity to apply the proposed method to the helthmonitoring of a practical composite structure.