Objective Aiming at the problems of low efficiency and inconsistent quality in intelligent forming process planning in the footwear industry, a bidirectional cyber-physical driven intelligent process planning method is proposed to promote the digital transformation of traditional shoemaking.

Methods A hybrid mechanism digital twin model incorporating geometry, physics, behavior, and rules was constructed, and a closed-loop feedback system for process planning based on bidirectional cyber-physical drive was designed. Furthermore, a forming process planning method driven by multi-feature coupling and real-time process feedback was developed and integrated into an intelligent forming management platform.

Results The bidirectional cyber-physical driven process planning system developed in this study has been comprehensively validated on an intelligent cold-adhesive shoe forming production line, with core achievements demonstrated across four dimensions: planning efficiency, product quality, resource consumption, and system intelligence. In terms of planning efficiency and responsiveness, compared to traditional process planning that relies on manual trial-and-error by engineers, the application of this system enables the digital twin model to automatically generate and simulate multiple optimization schemes, significantly reducing the average time required for planning the glue spraying path for a new shoe sole; meanwhile, for high-variety, small-batch orders, the production line changeover time is markedly shortened, greatly enhancing production flexibility. Regarding product quality consistency, through real-time inspection of the "glue line after upper-sole assembly" using vision sensors and feedback for compensatory calibration in the virtual model, the deviation of the upper glue line is controlled within ≤1.5 mm, significantly reducing the risk of debonding; the system also performs online inspection and control of "sole glue spray coverage", maintaining a stable coverage rate above 96%, thus avoiding quality issues caused by insufficient glue or overspray and leading to a substantial increase in the product qualification rate. In terms of production cost and resource consumption, precise glue spraying control reduces the overconsumption of adhesive, and the reliance on process engineers' experience is greatly diminished — the role of process debugging personnel for a single production line shifts from glue line debugging to monitoring and handling equipment exceptions, while the number of physical debugging iterations is reduced, thereby eliminating energy waste caused by equipment idling. Regarding process knowledge precipitation and decision intelligence, the system archives successful process parameters (e.g., robot trajectory, glue pressure, and temporal coordination) as digital assets, forming a "process knowledge base" that can automatically recommend optimal historical solutions when processing similar shoe styles, achieving a paradigm shift from "experience-driven" to "data and model-driven" operations.

Conclusion This research provides an effective intelligent process planning solution for the footwear industry, significantly enhancing production efficiency and quality stability.