Aiming at the problem of low production efficiency and inability to fully unleash production capacity in current semiconductor enterprises, a production efficiency optimization model for semiconductor enterprises has been studied and constructed. This model transforms the problem of low production efficiency into a problem of locating and solving the decoupling point of enterprise customer orders, and comprehensively considers the situation of sudden changes in enterprise production orders when locating and solving the decoupling point of customer orders. Propose to use a three-layer coding (TLC) mechanism to improve and optimize the biogeographical optimization algorithm, and use the improved biogeographical optimization (IOBO) algorithm to solve the production efficiency optimization problem of semiconductor enterprises. The results show that the proposed IBBO-TCL algorithm has a fast convergence speed and the minimum root mean square error after convergence. And this method can accurately solve the decoupling point of customer orders for semiconductor enterprises. The method proposed in the study has effectively improved the production efficiency of semiconductor enterprises and has guiding significance for optimizing enterprise structure.