This paper delves into the mechanical properties of Thermoplastic Vulcanizates (TPV) to address the issue of elasticity compared to traditional vulcanized rubber. TPV possesses unique advantages in the field of materials science due to its outstanding recyclability and excellent resilience, yet it still faces challenges regarding insufficient elasticity compared to rubber. This study proposes a comprehensive research approach by integrating Representative Volume Element (RVE) models and machine learning methods to comprehensively investigate the mechanical response of TPV. Machine learning models such as Support Vector Machine (SVM), Convolutional Neural Networks (CNN), and Random Forest algorithm (RF) are applied to predict the mechanical properties of TPV, and the accuracy of the models is validated. Through importance analysis and feature correlation analysis, the significant influence of the ratio of PP and EPDM's elastic modulus on TPV performance is revealed, along with corresponding adjustment theories. This research holds significant importance for understanding and predicting TPV material properties, providing profound insights for material design and engineering applications.