Abstract: We utilized the K2 algorithms and adaptive genetic algorithms to construct Bayesian network of college mathematics courses on the basis of 1984~2005’s 10 courses scores of school of statistics from a certain university. As far as the construction is concerned, it uses the information in the sample of students’ scores, combines Directed Acyclic Graph (DAG) with the theory of probability organically, intuitively reveals the dependencies of the courses and incarnates how strong thedependenciesare according to conditional probability table. Not only dose it offer the reference value on courses setting in different terms, but also it provides evidence for students’ advanced studies and option of research direction which is the most suitable.

Key words: Bayesian network, adaptive genetic algorithms, K2 alg orithms