The artificial bee colony algorithm (ABC), as a classic swarm intelligence algorithm, has advantages such as fewer parameters and clear logic. However, ABC cannot balance the exploration and development stages well in the iterative process, and is easily affected by local optimal solutions in the final optimization stage, which affects the final optimal solution. To effectively compensate for the shortcomings of the algorithm, a neighbor learning artificial bee colony algorithm based on permutation solutions (CNABC) is proposed. In CNABC, a dynamic neighbor learning strategy is proposed to improve the search ability and optimal selection ability of the algorithm in the exploration phase. To solve the problem of lack of balance between exploration and development, the local optimal solution is used to guide the update of the surrounding difference. After the three stages of the algorithm are completed, a substitution mechanism is introduced, which replaces the worst solution by introducing external candidate solutions as feasible solutions, thereby improving the algorithm’s ability to escape from local optima. Finally, comparative algorithm experiments are conducted on the CEC2022 test set, and compared with the TOP algorithm in the CEC competition on the CEC2022 test set. According to the experimental results, CNABC has good competitiveness in the comparative algorithm, which verifies the novelty and optimization ability of CNABC.