Abstract: To improve the performance of Artificial Bee Colony algorithm (ABC), an Improved ABC (IABC) for global optimization was proposed with the opposition-based initialization method. Inspired by particle swarm optimization algorithm and differential evolution algorithm, a new search mechanism was also developed to balance the exploration and exploitation abilities. The algorithms was applied to 4 benchmark function with effects of selective probability p. To verify the performance of IABC algorithm, 10 benchmark functions were tested with various dimensions. Numerical results demonstrated the proposed algorithms outperformed the ABC in global optimization problems.