The transfer function design is important for medical image visualization. Through the analysis of current transfer function design method, we propose a clustering algorithm with the help of gradation-gradient colored histogram to realize the classification for volume data, which is convenient to highlight the tissues of interest. Additionally, the number of these combined clusterings and the clustering centers can be determined rapidly through the color distribution of the gradation-gradient colored histogram. Of course, the method makes the clustering algorithm more effective and more reasonable. Currently, the transformation of three-dimensional data field from data attributes to optical attributes using the transfer function is carried out in the RGB color space. In this paper, combining the advantages of HSL color mode, we map the identifying information to HSL color space firstly. Through the method, it’s easy and intuitive for us to adjust the visual quality. Based on the method of transfer function design for volume rendering given in this paper, we conduct an experiment with the program of OpenGL/GLSL to realize the medical image visualization using ray-casting volume rendering algorithm. The experimental results show that this method is effective and corrective.