As Internet use has proliferated, web-based learning systems have become more and more popular. Numerous researchers have spent a great deal of effort to facilitate the promotion of high quality web-based learning environments, such as intelligent web-based learning systems and adaptive learning. To facilitate such researches, students’ behavioral patterns must be observed and experimentally analyzed. However, building a web-based learning system and the requisite collecting of behavioral patterns usually takes a great deal of time and effort. To solve this problem, this paper proposes a learning behavioral model based on Colored Petri Nets (CPN) to model and generate students’ behavioral patterns. To verify the viability of the proposed model, this paper compares actual data collected from elementary school students with the behavioral pattern generated by the proposed model. The results prove: (1) The generated behavioral pattern approaches actual student behavior; (2) The generated behavioral pattern serves as adequate test data to test whether the predicted learning content of an intelligent e-learning system is appropriate; and (3) The proposed model is capable of recommending the appropriate learning content for students utilizing e-learning systems.