In this study, model based damage identification algorithms were applied for a delaminated composite beam and plate. Initially, natural frequencies of healthy and delaminated beam were measured via experimental modal analysis. Next, the natural frequencies obtained by experimental method were verified with the finite element analysis. Once the validation has been fulfilled, the next step was to apply the finite element analysis to extract the mode shapes. Then, the curvatures mode shapes were obtained from the mode shapes, and these parameters were used in the gapped smooth method (GSM) and strain energy method (SEM) to detect the location and size of the delamination. The damage diagnosis technique for composite plate was evaluated for two cases. In the first case, it was assumed that the initial health information of a structure is known. However, the information of an undamaged structure may not be available during the diagnosis process for some circumstances. Therefore, the diagnosis was conducted by evaluating the curvatures of damaged structure using surface fit techniques in order to predict the initial status for the second case. Both of the two damage detection algorithms were evaluated for these two cases. The results show that, the delamination of laminated composite beam and plate could be predicted non-destructively by employing the data obtained from experimental or numerical technique in model based SHM approach for laboratory conditions.