This paper proposes a new image segmentation method which uses 3D color histogram and human visual attention analysis. Existing histogram-based methods find the cluster centers by analyzing the distribution of the 3D color histogram values. It is difficult to consider the overall characteristics of an image by analyzing the distribution of the 3D color histogram values. Thus, Existing histogram-based methods have difficulty to finding the optimal cluster centers. To overcome this drawback, the proposed method uses both 3D color histogram and human visual attention analysis to find the optimal cluster centers by considering the overall characteristic of the image. Compared with the benchmark methods, the experimental results show that the proposed method improved the segmentation accuracy.