Sequenced personalised curriculum is an important research issue for Web-based educational systems, because it does not uses the fixed learning paths approach, as it will not be suitable for all learners. This paper presents a framework for Adaptive-Recursive Tutoring System (ARTS) that classifies the learners and provides cognitive adaptation according to their learning profiles and abilities. ARTS support the use of knowledge-based systems. In particular, the adaptive-recursive learning and teaching techniques, this will help deliver flexible and assessable course materials in education or business sectors. ARTS recognise the importance of individual learner knowledge. ARTS adaptive agent customises courseware and coaches the students through the learning journey. ARTS is equipped with a recursive agent (collectively with the adaptive agent are called the instructor Model) that call itself on a smaller version to augment and monitor the student learning process.