Curriculum modeling is a critical component in the development of adaptive learning systems, as it defines the structure and sequence of learning objectives, enabling adaptive sequencing of content and activities. While significant research has been devoted to domain and student models, curriculum modeling has received comparatively less attention in the context of adaptive learning. In this paper, we propose an approach to curriculum modeling designed to enhance adaptive learning. Our approach incorporates granular structure, outcome-based education, and graph theory, based on curriculum mapping, mastery learning, and human-AI collaborative decision making. This approach's feasibility has been shown through its implementation in an adaptive formative assessment system.