Aiming at the disadvantages of the imbalance between exploration and exploitation in the artificial bee colony algorithm, as well as the resulting low solution quality and slow convergence speed, this paper proposes an artificial bee colony algorithm based on stimulus-response labor division. Firstly, exploration and exploitation are regarded as two different search tasks, bees explore the search space in the employed bee phase and exploit the search space in the onlooker bee phase. Then, the environmental stimulus of the search task is designed according to the population diversity, and the response threshold of bees is designed according to the search success rate. Finally, under the mechanism of stimulus-response labor division, bees flexibly switch between the employed bee and the onlooker bee to achieve the exploration-exploitation balance. The experimental results on 22 benchmark functions show the high efficiency of the proposed algorithm.