Disassembly lines have to face task-failure situations due to the variability in quality of incoming product. Such failure violates the precedence relationship for the remaining task at downstream stations. Therefore, task failure requires corrective measure to improve the profitability of disassembly line. In this paper, a recursive optimization approach has been proposed to improve the profitability of disassembly lines, which takes corrective measure to determine optimal sequence of tasks. For this purpose, Teaching Learning Based Optimization (TLBO) algorithm has been used to find optimal sequences before and after task failure. To reduce the computational time required during recursion, the proposed solution approach is equipped with memoized list for finding corrective measure. A numerical illustration has been used to demonstrate the applicability of proposed solution approach which is capable to handle high variability in quality of incoming products.