Abstract: In view of the uncertainties and the complexity of parts in the disassembly line, we build a multi-objective mathematics model that includes minimizing the number of workstations, idle index, disassembly cost, and the classification index and present an improved fireworks algorithm to solve it. First, we discretize the fireworks algorithm considering the characteristics of the problem to be solved and redefine the explosion operation and mutation operation. After the explosion and mutation sparks are generated by the fireworks individual, we use Pareto solution set and NSGA-Ⅱ congestion distance mechanism to screen the feasible solution and update individual fireworks. Second, we use the proposed fireworks algorithm on the disassembly-line balancing problem of linear medium-scale and U-shaped large-scale examples and verify the performance of the improved fireworks algorithm by comparing it with other algorithms. Finally, we apply the model and algorithm to the printer disassembly line. The proposed mathematics model and algorithm are more effective and feasible than linear solutions.