Abstract: Artificial bee colony (ABC) algorithm is a novel search algorithm which simulates the intelligent foraging behavior of honeybee swarm to solve the practical problems. However, there is only a behavior communication way (dancing) in the current ABC algorithm, which results in the lack and lag of collaboration among bees and influences the solving performance of ABC algorithm. Inspired by the objective fact of transinformation among real bees, a new ABC algorithm is proposed by introducing a chemical communication way based on inductive pheromone and applied to solve multidimensional knapsack problems (MKP), which is more faithful to the transmission information of real bee colony system. With the combination of the behavior communication way and the chemical communication way, the new algorithm makes the honeybees cooperate with each other better by the scheme of inductive pheromone updating and diffusion. A number of simulation experiments and comparisons on benchmark datasets of MKP demonstrate that the performance of the new algorithm is superior over the original ABC algorithm. The performances of the new algorithm have also been compared with some typical meta-heuristic search algorithms, and the computational results show that the new ABC algorithm obtains better quality solutions than all the other approaches.