Our study is concerned with making the instruction suitable to the individual learner’s characteristics. This chapter describes the methodology used to investigate how to model the learner’s Cognitive Ability Level (CAL) based on the observation and analysis of his/her behaviour in a Web-learning environment. In our study, the CAL represents the learner’s cognitive stage development according to Bloom’s taxonomy. The methodology encompasses two phases: (i) the generation of the CAL classes for the target population and (ii) the study of learning trajectories of CAL classes in an experimental learning module. As a result, we have identified the CAL classes’ parametersvalues that best discriminate these classes from the observation and analysis of their learning trajectory on the Web. The entire knowledge obtained from this investigation will make possible to automate the learners’ CAL diagnostic. It will also give us the background to develop Web-learning environment contents.