In order to segment images exactly and quickly, based on the traditional adjustment of colony algorithm selection strategy and defect honey, a new method based on a improved Artificial Bee Colony algorithm segmenting two dimensional Otsu images is proposed. This method looked on the image threshold value as artificial colony algorithm of the bees. The best threshold is approached in parallel via the division of labor, cooperation and information sharing of employed bees, onlookers and scouts. Effectively solved the problem of the traditional two dimensional Otsu image segmentation calculation defects, long operation time. Experimental results show that the proposed algorithm not only can get the ideal segmentation results, but only improved the segmentation speed.