In this paper, Acoustic Emission (AE) and principal component analysis (PCA) combined with fuzzy prediction method were used to establish a failure prediction model of polyamide (PA) plastic, and the relationship between AE signal and material failure degree during PA stretching was studied. The constant speed tensile failure experiment of PA material was carried out at four different tensile speeds, and AE signals during the tensile failure process of PA plastic were monitored and collected in real time. The evaluation factors were determined as amplitude, ringing count and rise time through correlation coefficient analysis. The triangular membership function determined the failure membership degree of each review set. The results showed that the combined fuzzy prediction method of AE and principal component analysis (PCA) could monitor the damage evolution of material tensile process and provide a reference for the health monitoring of PA materials. The model could resist noise or other interference and minimize the uncertainty caused by human factors. Thus, it could enhance the accuracy and credibility of the analysis results greatly, and improve the reliability of PA damage degree prediction efficiently.