Understanding play traces resulting from the learner's activity in serious games is a challenged research area when the serious game allows a large amount of free actions. In this paper, we present a technical and a methodological framework that assists designers to build Petri nets of a serious game and then use them to analyze player's traces and generate pedagogical labels about the learner's behavior. We carried out an experimental study to evaluate the effectiveness of the labeling algorithm and to validate the acceptation and the readability of the pedagogical labels by the teachers.L'analyse des traces des apprenants dans les jeux sérieux est un défi lorsque le jeu sérieux autorise une liberté d’action importante au joueur. Dans cet ...