Researches into E-Learning pay emphasis on how the technology is applied for comprehensive learning support in various perspectives. Lots of advantages are involved while performing the E-Learning. For example, timely contents are applicable to be delivered via the Internet, and in a similar manner, instant feedback which identifies the learning performance would be obtained through an online assessment system. In this article, a diagnostic approach for learning performance evaluation of student is proposed. Two primary issues are concentrated based on the analysis of quantitative assessment results. First the authors identify the misconceptions on specific subject(s) that students have learned and provide remedial suggestions about their learning status to students. Second, because questions may have several incorrect information to students, hence the authors approach provide suggested revisions to instructors. The proposed approach employs the Rule-Space model and the Student-Problem chart and is implemented in an online assessment system to obtain corresponding data (e.g., learning behavior) and performance (e.g., learning status). The feasibility is demonstrated through an empirical study (i.e., a series of pre-post tests) which was conducted in a domain-specific course (i.e., Fundamental of Information Technology) with around 450 students in a calendar year.