Aiming at the segmentation of liver region in the abdomen CT image, we propose a novel method based on Grow-Cut algorithm. The Grow-Cut is an iterative algorithm, as the automation labels the image, user can observe the segmentation evolution and guide the algorithm with human input where the segmentation is difficult to compute. In this paper, we propose a new energy function for the traditional Grow-Cut method to obtain the better segmentation precision. Moreover, we do a pretreatment using the k-means to decrease the running time. In additions, we take multi labels for the Grow-Cut to get multiple organ segmentation results in one operation. Lastly, we take several experiments to demonstrate the validation of our proposed algorithm, and experimental results show that our method has a good robustness.