A computational inverse technique is presented for characterizing the material property of functionally graded material (FGM) plates, using the dynamic displacement response on the surface of the plate as input data. A modified hybrid numerical method is used as the forward solver to calculate the dynamic displacement response on the surface of the plate for given material property varying continuously in the thickness direction. A uniform crossover micro-genetic algorithm (uniform uGA) is employed as the inverse operator to determine the distribution of the material property in the thickness direction of the FGM plate. Examples are presented to demonstrate this inverse technique for material characterization of FGM plates. The sensitivity and stability to noise contamination in the input displacement response data are also investigated in detail. It is found that the present inverse procedure is very robust for determining the material property distribution in the thickness direction of FGM plates.