Aiming at the complexity of technics and the problems on arranging production processes wholly by hand in an iron and steel industry, a three-tier production planning and scheduling model was proposed. Contracts were grouped into virtual contracts through matching virtual production numbers got by collecting product attributes. The model including general month production planning, plant monthly production planning and plant daily production planning, and scheduling model are presented to deal with virtual contracts. The plant daily production planning is represented logically by binary search tree and rule-based knowledge base is adopted to work out scheduling model step by step. Finally, a production planning module is developed based on J2EE for the steel industry and the proposed approach is demonstrated through a real application to the steel industry