Globally, Industry 4.0 technologies' rapid development advances the rising role of design-led integrated STEM education in schools. However, the absence of problem-finding abilities among students may make it difficult to carry out the educational program. Adaptive learning with artificial intelligence (AI) can be used to construct and implement a learner-centred intelligent tutoring system, which will enable students to adapt and engage in relevant learning tasks by improving their transdisciplinary problem-finding skills and stocking authentic contextual information designs. In order to solve the research problem, this study explores three significant areas of information: 1) automatic construction of knowledge structures; 2) individual ability values and group classification; 3) adaptive recommendation of review content and assessment tasks. As a result, this study promotes the development of an intelligent tutoring technology framework containing data, algorithms, and services for integrated STEM, thereby enhancing a social atmosphere that values science and encourages innovation.