The knowledge grid (KG) has gradually been becoming a keen research topic in the joint field of the Semantic Web, the Semantic Grid, and the Web Services for its potential impact in making full use of various resources, being it conceptual or physical, from the web and grid. Many research activities in this direction focus on the architecture and its components, and base themselves technically on data mining technologies or the like, which obviously considers (web) information systems to be a foundational layer for the grid and web. However, further study on the web has revealed that the web cannot, and unnecessarily, be viewed simply as a closed world model (CWM). The web is an incomplete and inaccurate (knowledge) system, from which we will seek answers regardless their inconsistency. In this paper, we will first investigate the current researches into the knowledge grid, then discuss the problems of incompleteness of knowledge processing and reasoning and propose a category theory based model to handle such incomplete knowledge, and finally provide an analysis of the model.