With the rapid increase of large-scale distributed manufacturing projects and global outsourcing, it is necessary for manufacturers to have a uniform Web-based environment to ensure the quality of products. This paper discusses a graph–based semantic data model using the Category Theory for a knowledge–based Virtual Surface Texture Analysis (VSTA) system, which aims to automatically implement surface texture parameters relating to surface characteristics defined in the Geometric Product Specification and Verification standards (GPS). The paper starts with a literature review on related research areas such as Virtual Environment, Virtual Manufacturing and GPS and a brief introduction on the background of this project. Based on the problems highlighted, a categorical data modeling method will be introduced and its usage explained. A web-based virtual environment for hosting the VSTA system interface will also be discussed. The paper concludes with a case study of the implementation for the VSTA system to illustrate the design features of the devised categorical data model.