Abstract: An adaptive mixed culture artificial bee colony algorithm (AMC-ABC) is proposed to solve continuous space optimization problem. In the algorithm, community space is evolved by the improved group update way with optimal foraging theory; the knowledge of belief space is updated by the cloud model algorithm and optimal sorting differential mutation strategy; the outer space is evolved by chaos algorithm and opposition-based learning algorithm; and the knowledge exchange of three kinds of spatial is realized by adaptive acceptance operation and effect of operation. Simulation results of the typical complex functions show that the algorithm has fine the convergence precision and computing speed, particularly suitable for optimization the multimodal function.