Thresholding is an important form of image segmentation and is used in image processing and recognition for many applications. In this paper, an automatic approach for thresholding based on gray level gradient co occurrence matrix model and the maximum entropy principles is proposed. This method utilizes the information of both gray level and gradient in an image. In this approach, the threshold vector is selected through evaluating two dimensional entropies based on the gray-level gradient co-occurrence matrix and maximizing the edge region entropies. It is found that the proposed approach performs better than other 2D entropy methods.