Over time, various learning management systems have been developed. One of these is MEITREX, a gamified intelligent tutoring system, which has been developed specifically for software engineering eduction at higher education institutions. By providing students with individual feedback and learning material adapted based on students’ individual process MEITREX should increase students motivation. To provide students with individual feedback and to adapt learning materials based on students’ current knowledge status, the student’s current knowledge status needs to be automatically and reliably determined. Currently, MEITREX uses a simple score to determine students current knowledge. This approach has several drawbacks. This approach for example doesn’t determine students’ knowledge of a single skill, but of the content of a chapter. Additionally, a student needs to repeat each assessment of a chapter a certain number of times to master the content of a chapter, despite having mastered the content before. Over the past decades, different approaches for the estimation of students’ knowledge status have been introduced. The introduced approaches range from simple machine learning models to complicated neural networks, hidden markov models and approaches, that have their origins in the chess world but have been modified. All approaches have in common, that their estimation of students’ knowledge status is based on the students’ performance on exercises. One of these established and reliable approaches should replace the old score. Not all models are equally well suited for usage in an MEITREX. Therefore the requirements such a model needs to meet are defined and a requirement analysis based on the existing literature is conducted to find promising model groups, as the existing approaches can be grouped into eight groups of models. Based on the results of this requirement analysis the performance of eight promising models from three different model groups is tested. Of all tested models, that fulfill the most important requirements M-Elo showed the best performance and therefore M-Elo was integrated into MEITREX. Junit Tests and a short evaluation showed, that the integration M-Elo into MEITREX was successful.