Magnesium alloy is the lightest material in all practical metals, and has the excellent mechanical properties. From such reasons, the alloy is noticed as the attractive material in various fields. However, the alloy has the weak points that are inferior for heat and corrosion resistance. Therefore, it will be possible to guarantee the safety security of structure parts, if the information on the fracture strain and life, etc. are evaluated by the signals detected from the alloy under those use environment. The three type specimens of the virgin, the heat history (i.e. repeatedly gave the heating to the virgin specimen) and the corrosion specimen (i.e. immersed the virgin specimen into solution) are prepared in this study. Then, the tensile test was carried out for each specimen, and the AE signals emitted in the testing were detected. Frequency and fractal dimension analysis were carried out in order to extract the features of those signals. From such analysis results, the feature extraction of the signal by the elasticity twin deformation was done, because that deformation participates in the degradation of mechanical properties. As the result, the following things became clear under the experimental condition in this study: The frequency component of the signals caused by that deformation is near 550 kHz. The fractal dimensions Fd of those signal show the value over 1.85. Then, it was examined whether the fracture strain is evaluated by the Ib value (Improved b-value) led from the frequency distribution of the Fd. As the result, the effectiveness of this method proposed under the experimental conditions in this study was confirmed. Therefore, the possibility for the prediction of that strain by the AE signal detected in elastic region was shown.