<![CDATA[he transformation of learning through the Deep Learning approach requires Islamic Religious Education (IRE) to move beyond content mastery toward fostering students' religious moderation. This study aims to analyze and develop a conceptual model of adaptive Islamic Religious Education learning orchestration through a Deep Learning approach grounded in religious moderation in madrasahs. A descriptive qualitative design was employed. Data were collected through classroom observations, in-depth interviews, and document analysis, and were analyzed using the interactive model of Miles, Huberman, and Saldaña. Data trustworthiness was ensured through source and method triangulation, member checking, and prolonged engagement. The findings reveal that adaptive IRE learning orchestration is constructed through four interrelated components: adaptive instructional planning, meaningful, mindful, and joyful learning interactions, the internalization of religious moderation through authentic learning experiences, and continuous adaptive assessment. The synthesis of these components resulted in an Adaptive Islamic Religious Education Learning Orchestration Model, positioning teachers as learning orchestrators who integrate instructional design, learning experiences, character development, and assessment into a coherent learning system. This model extends the Deep Learning perspective by embedding religious moderation as the core principle of adaptive and contextually relevant Islamic Religious Education in madrasahs.]]>
