The paper describes a 3D Craniofacial visualization system developed for the high precision medical robotic system. An open framework is designed to improve the scalability of the platform, by which new features can be easily plugged in the platform. A variety of efficient tools, such as stoke-based direct volume object cutout, are provided to reduce the tedious user interaction in the surgical design process. A number of strategies are adopted to improve the quality of 3D medical visualization, including optimal transfer function design based on one-dimensional/high-dimensional functions, automatic/semiautomatic volume classification, multiple feature highlighting, feature variation curve analysis. Meanwhile, “shader assembly line” supports a scalable mechanism for new visualization effects by restructuring visualization algorithms based on accessories-like elements. Experimental results show the effectiveness of our system.