Segmentation is an important application in the analysis of gray level and color images numerically. The task of image segmentation is to separate a given image into different regions having homogenous properties. Although the most basic segmentation is thresholding, this method has been insufficient in time. In this paper, the segmentation of color texture images is carried out with the approach of wavelet transform and artificial bee colony algorithm. Through this approach, the segmentation of typical natural scenes, including no uniform statistical color and texture characteristics is achieved. The superior and weak points of the developed method are evaluated.