Abstract: A novel algorithm is proposed to segment the tumor region of human liver CT image. Firstly, the region of interest (ROI) is segmented using the dynamic adaptive region growing method. Then the wormhole of the ROI is filled by morphological filtering to get the tumor region. Groups of CT images are used to test the efficiency of our algorithm. The experiments show that the algorithm is very useful for segmenting the tumor region of liver CT images. It is extremely important for calculating tumor volume and clinical treatment of the liver tumor.