This paper presents a systematic approach to the disassembly line (DL) design in meeting the requirement of variant orders for multiple used parts with different due dates. An extended disassembly Petri net model is proposed for the hierarchical modeling in order to derive the disassembly path with the maximal benefit in the presence of some defective components. An algorithm for balancing DLs to maximize the productivity of a disassembly system is presented. The results of simulation runs of the proposed methodology and algorithms applied to a simplified personal computer disassembly are provided. This work lays a foundation for designing efficient industrial automatic and semiautomatic disassembly systems. Note to Practitioners-Disassembly is rapidly growing in importance as manufacturers face increasing pressure to deal with obsolete products in an environmentally responsible and economically sound manner. This process can be performed at a single workstation or on a disassembly line (DL) that is organized as a sequence of workstations, each with one or more machines/operators to handle a certain type of disassembly task. Compared to a single workstation, DL provides higher productivity and greater potential for disassembly automation. However, it still faces serious scheduling and inventory problems because of a high degree of uncertainty in discarded products and disparity between demands for certain parts and their yield from disassembly. To address this challenge, this paper proposes a two-level systematic approach, aiming to maximize system throughput and system revenue by dynamically configuring the disassembly system into many DLs, while considering line balance, different process flows, and meeting different order due dates. The research results can help engineers build better disassembly systems.