Aiming at the low search efficiency of the traditional backtracking algorithm when planning the function sequence of the payload system based on the knowledge model, an improved algorithm named as WIM-CGA for CGA (co-evolutionary genetic algorithm) was proposed, which was based on the WIM (worst individual mutation) strategy. The algorithm adopted a dual-route evolution scheme in the genetic process, which was “the better individuals perform standard genetic processes, and the worse individuals perform mutation operation”, to improve the solution accuracy and search efficiency. Simulation results show that under the same test conditions, when the function scale is 50 and the constraint density is 1.0, the average accuracy of the optimal solution of WIM-CGA within the limited time is 54.15% higher than that of GAC-BS (BS based on generalized arc consistency) and 6.18% higher than CGA, and when optimal solution accuracy reaches 90%, the iteration times of WIM-CGA is 65.79% lower than that of CGA, and the time consumed is reduced by 48.97%. The efficiency of functional sequence planning is improved significantly.