Computer applications in education are now common to complement a classroom teaching and learning activity. However, a reportedly obscure area in the computer-based instruction is how it can systematically implement the different learning path for the diverse levels of students. In this respect, the dynamic (or adaptive) curriculum sequencing is quick to admit its advantages, and shows the best way forward in the computer-based instruction, matching the content with each individual's learning performance. This article further discusses this matching process with the `flow' theory, by which one can draw upon a way to find an optimal learning path in Intelligent Tutoring Systems (ITSs). A computer-based learning system - `IT-Tutor' was thus implemented and used to empirically investigate this issue further. The results of the study suggested that key to dynamic curriculum sequencing might be setting out optimal learning experience, in conjunction with both the knowledge level and challenge level of each individual.