With the advance of Internet technology, learning on-line has become one of the most common as well as popular ways to knowledge. However, in the environment of Asynchronous distance learning, it is difficult for teachers to have in hands all the students’ individual learning conditions and unrevealed information without face-to-face communication, especially when the numbers of students increase day by day. With the focus on student assessment model analyzed with Bayesian method, and the combining of graphic structure, this paper attempts to set up a system to model students’ assessment information, providing graphic vision to teachers, and moreover analyze and estimate other students’ assessment information (or their other assessment information). In short, this paper aims to build up a student model based on student portfolio, which can further apply to individual homework assignments, diagnoses of students’ learning problems, inference of students’ learning condition, and thus promotes teachers’ teaching performance.