<p>The rapid progression of machine learning (ML) has revolutionised numerous fields, including engineering, where these technologies are being leveraged to optimise design, improve efficiency, and automate complex processes. In next-generation polymer composites (NGPC), ML is driving a paradigm shift in materials science, offering new opportunities to transform how composites are designed, manufactured, and tested. These tools enable the development of high-performance, cost-effective materials by predicting optimal material combinations, fibre orientations, and structural configurations, significantly reducing the reliance on traditional trial-and-error methods. This paper provides a comprehensive review of the current state-of-the-art ML applications in NGPC, focusing on key areas such as material engineering and selection, optimisation and modelling of manufacturing processes, and the prediction of material properties. Furthermore, it underscores the role of cutting-edge ML techniques in damage assessment through non-destructive testing and structural health monitoring of composite structures. Despite these promising developments, ML applications in NGPC remain in a relatively early stage, with ongoing efforts needed to overcome limitations in data availability, model generalisability, and practical deployment. The review concludes by outlining current challenges and future research opportunities for integrating modern ML approaches into NGPC, offering valuable insights for researchers and engineers in this rapidly evolving domain.</p>