Shear wave elastography (SWE) is a powerful method for the diagnosis of tissue disorders or degeneration based on tissue elasticity. In SWE application, it was recognized that the wave speed depends not only on the tissue elasticity but also on the structural shape, leading to different theoretical models. For liver, skin and myocardium, the appropriate theoretical model is known to be shear wave, Rayleigh wave and Lamb wave theory, respectively. Therefore, appropriate theoretical model should be adopted for the proper application of SWE. In this study, we verify these theoretical models in gelatin samples of different thicknesses, using experimental and numerical SWE tests. The results indicate that the wave speed was influenced by the ratio of the wavelength and sample thickness and the measurement region. Based on these results, the selection of theoretical model could be divided into three cases, and the appropriate theoretical model can be selected accordingly.