Abstract: In wide and heavy plate production, surplus materials present with diverse specifications and complex states, while manual grouping-plate suffers from low efficiency and poor stability. To address these issues, an intelligent grouping-plate system for surplus plates was designed, developed and put into online application. The system focused on multi-constraint matching between surplus plates and orders. A feasible matching model based on metallurgical and process rules was constructed, followed by a multi-objective collaborative optimization model that comprehensively considered material yield, inventory time, delivery requirements, processing costs, and economic benefits. To handle the large-scale and real-time requirements, a heuristic intelligent grouping-plate algorithm incorporating domain-specific rules was designed to achieve rapid and optimal matching decisions between surplus materials and orders. The system has been successfully applied in multiple typical wide and heavy plate production lines. Practical operation results demonstrate that, compared with manual operation, this system can process surplus materials in real time. The volume of surplus materials automatically processed is, on average, more than 10% higher than that achieved manually, while the comprehensive utilization rate of surplus materials is effectively increased by more than 2%. This serves as a good demonstration for improving the yield and promoting the refined management of surplus materials in steel enterprises.