The artificial bee colony algorithm (ABC) is a kind of stochastic optimization algorithm, which is used to solve optimization problems. In view of the shortcomings of basic ABC with slow convergence and easily falling into local optimum, a modified artificial bee colony algorithm (MABC) is proposed. First, a high dimension chaotic system is employed for the sake of improving the population diversity and enhancing the global search ability of the algorithm when the initial population is produced and scout bee stage. Second, a new search equation is proposed based on the differential evolution (DE) algorithm, which is guided by the optimal solution in the next generation of search direction to improve the local search. Finally, a learning probability (P) method is introduced, corresponding to different value with each particle. Thus, the capacity of the exploration and exploitation of each particle in the population is different, which can solve different types of problems. The performance of proposed approach was examined on well-known 10 benchmark functions, and results are compared with basic ABC and other ABCs. As documented in the experimental results, the proposed approach is very effective in solving benchmark functions, and is successful in terms of solution quality and convergence to global optimum.