Abstract The Disassembly Line Balancing Problem (DLBP) is of significant importance in the product recycling process. However, existing DLBP research have primarily proposed improved optimization algorithms for single solution space. To provide decision makers with more efficient disassembly solutions, this paper chooses three classical layouts (straight, U-shaped, and parallel) to broaden the solution space and expands an incomplete DLBP with multi-solution spaces (DLBP-MS). By employing the strategy of incomplete disassembly, only the essential components are retained. In DLBP-MS, the disassembly information from used products is processed and imported into the three types of disassembly spaces to be solved and compared to find suitable disassembly solutions. And a multi-objective mathematical model is developed, this includes factors such as workstation count, free time, disassembly smoothness index, carbon emission, and disassembly revenue. The part constraints are established based on directed graphs, and the encoding and decoding methods for multi-solution spaces disassembly sequences in the random incomplete case are designed respectively. A ring topology based flower pollination algorithm (RTFPA) is introduced to effectively address the DLBP-MS. The solution set obtained during the iterative process is divided into subsets based on the congestion level and the overall diversity is preserved by internal optimization of these subsets. After that, the RTFPA is applied to the DLBP-MS of waste cell phones and computers. The improvement of the algorithm's optimization ability under different solution spaces is verified by comparing the results with those obtained from four other algorithms.