The existing online education system mostly provides users with undistinguished learning content. The existing online education system assumes that users have the same characteristics or tendencies as one another. Every user is not the same; they have different characteristics, tendencies, and learning preferences. Each user has different personal characteristics, but today's online education systems ignore them. This study use Felder Silverman's learning style model. Through characteristic learning styles,relationships, and content types result in more personalized e-learning. Researchers distributed questionnaires to Informatics Engineering students at, University of August 17, Surabaya, and managed to get 277 student respondents. This data uses to determine student learning styles using a questionnaire-based psychometric instrument called the index of learning style (ILS), which uses to determine learning styles. The questionnaire, validity, reliability, determination of learning styles, learning content, and material requirements are testing. This study concludes that the questionnaire used meets the standards for determining learning styles for users. The system that builds has five modules to facilitate construction. These modules include the media library, domain repository, student model, instruction model, adaptive engine, and user interface. It can conclude that the modules made can help lecturers to distribute material according to the material content that is suitable for student learning styles and helps students to study more effectively according to their preferences.