In this paper, an architecture is proposed for the development of an e-Learning system based on Petri Nets and on a hybrid LDAP-SQL database. In this approach, e-Learning paths are initially tailored to the userpsilas aims and, during the learning process, are dynamically adapted to his needs. By means of an ad hoc user-system interface, each learner is allowed to specify his initial target and, if he decides to deepen some issues, is authorised to redefine such target over time. This process is guided by a ldquoPath Decision Modulerdquo based on Petri Nets, which checks the userpsilas learning level and consequently decides whether further issues can be faced or not. The underlying LDAP-SQL database benefits from the joint use of directory services and relational databases, used for the storage of learning material and dynamical paths respectively. Although the focus is on the e-Learning module, this architecture can be considered for a possible integration with m-Learning facilities.