Comprehensive review of WPC pyrolysis mechanisms and catalysis with emerging AI opportunities

Wood–plastic composites (WPCs), as hybrid materials of lignocellulosic fibers and thermoplastics, present significant challenges in end-of-life management due to their multiphase and multicomponent nature. Pyrolysis has emerged as an effective thermochemical strategy for converting waste WPCs into fuels and value-added carbon products. This review systematically summarizes recent advances in WPC pyrolysis mechanisms, catalytic upgrading, and kinetic modelling, with particular emphasis on how machine learning (ML) and artificial intelligence (AI) are beginning to reshape process understanding and optimization. The review discusses the catalytic roles of zeolites, metal oxides, and activated carbons, as well as the influence of fillers, compatibilizers, and ageing effects on product selectivity. Increasing attention is being directed toward hybrid kinetic–AI frameworks that integrate experimental datasets with mechanistic constraints to achieve interpretable, transferable, and data-efficient modelling of complex biomass–plastic interactions. Beyond methodological advances, AI-enabled analytics offer practical benefits including accelerated experimental design, improved uncertainty quantification, and data-driven guidance for scale-up. Finally, future priorities are highlighted, including standardized data infrastructures, multi-scale kinetic–reactor modelling, and integration with techno-economic and life-cycle assessment to support carbon-negative and industry-ready WPC pyrolysis systems. • Summarizes recent progress in catalytic and non-catalytic pyrolysis of WPC. • Reviews kinetic modelling methods and hybrid machine learning frameworks for WPC pyrolysis. • Discusses the influence of catalysts, additives, and aging on product yield and composition. • Identifies data standardization and model transferability as key challenges for AI integration. • Proposes an AI–mechanism coupled framework toward intelligent and carbon-negative WPC valorization.