Polymer blends have attracted significant attention in materials science in recent years [1]. In polymer design, blending polymers is a cost-effective and efficient strategy to tailor material properties for engineering applications. In polymer recycling, polymer blends are prevalent, as attaining 100% pure plastic fractions is often impractical. However, the presence of multiple polymers in a blend can significantly change mechanical properties [2]. These compositional variations introduce complexities that challenge the accuracy of existing constitutive models, thereby limiting their potential for reuse and high-value applications. To address this challenge, this study proposes an efficient and cost-effective hybrid FEM-NN approach for modeling the nonlinear mechanical response of pure polymers and their blends under various loading conditions. This approach integrates an NN into the constitutive model to effectively predict a key internal variable ruling the plastic kinematics—athermal shear resistance—which is highly sensitive to temperature and strain rate. The loading conditions, composition of polymer blends, and instantaneous strain can be considered as inputs to fully capture their influence on the mechanical properties. The NN-based constitutive model is implemented within a classical FE framework to conduct structural analyses. This hybrid approach is not restricted to a specific polymer or limited loading conditions and it can be seamlessly extended to a wide range of polymers and blends, making it a robust tool for material modeling and analysis. [1] Jagadeesh, P., et al. Sustainable recycling technologies for thermoplastic polymers and their composites: A review of the state of the art. Polym. Compos., Vol. 43, pp. 5831-5862, 2022. [2] Shan G. F., et al. Mechanical properties and morphology of LDPE/PP blends, J. Macromol. Sci. Part B-Phys., Vol. 46, pp. 963-974, 2007.