Evaluation is a critical and a "must-do" step in every e-learning course since it provides valuable feedback in order to tailor, redesign, improve or accept the course according to the stakeholders' needs. Evaluation by its nature is a complex process and constitutes an ill-structured problem. A series of critical factors are involved such as: a considerable number of different stakeholders (students, tutors, e-learning designers, developers, and so on), the functionality and usability of the available learning technology, the pedagogical approach (learning theories, instructional design models, learning strategies), the delivery mode, constraints (time, budget, people). Moreover, the outcomes of e-learning are difficult to be measured due to time and quality borders. A major goal of an evaluation process is to measure the underlying set of quality factors as perceived by the target stakeholders (e.g., students). The analysis of stakeholders' preferences is an essential factor to the success of the e-learning course because it gives valuable feedback to the decision makers in order to properly redesign/improve the course. In this chapter, we describe in detail the results of a new methodological approach for the analysis of students' preferences in e-learning courses based on the principles of theory of committees and elections and multicriteria disaggregation-aggregation decision aid approaches. The underlying real case study concerns the analysis of the students' preferences of an undergraduate course at TEI Piraeus during the academic year 2009- 2010. Multicriteria disaggregation - aggregation approach for discrete alternative actions aims to the assessment of an additive preference value model triggered by students' global preferences on a set of alternative actions, evaluated on a consistent family of criteria. Theory of committees and elections provide the way to structure acceptable global ranking from individual expressed preferences. The case study shows an in depth analysis of students' preference models along with an effective identification of the underlying changes in order to improve the whole course. Also, our new methodological approach seems very promising to be used in a wider and multipurpose evaluation structure (e.g., course, program, curricula).