Mathematical problem-solving requires learners to interpret task conditions, select appropriate strategies, monitor solution processes, and evaluate the reasonableness of answers. However, classroom instruction often emphasizes procedural accuracy without sufficiently making learners’ thinking processes explicit. This design-and-development study developed a theory-based metacognitive instructional design for Grade 6 mathematical problem-solving, specifically in finding the area of composite figures and solving routine and non-routine problems involving composite figures. The study translated Divinagracia’s Metacognitive Theory into a classroom-ready instructional sequence using four development phases: contextual analysis, theory-to-practice mapping, expert review and revision, and implementation planning. The design was anchored on seven metacognitive domains: knowledge of the typology of mathematical problems, knowledge of the nature of mathematical problems, awareness of mathematical knowledge and thinking, knowledge of personal strengths, knowledge of problem-solving emotions and attitudes, knowledge of thinking associated with bodily motion experiences, and metacognitive solution qualities. Expert review informed revisions that simplified learner prompts, improved time allocation, reduced cognitive load, and embedded metacognitive processes through guided questioning and observable learner actions. The resulting design integrated shape recognition, routine and non-routine problem analysis, guided formula application, self-questioning, hands-on measurement, real-life valuing, and structured reflection. The study contributes a context-sensitive instructional model for operationalizing metacognition in elementary mathematics by aligning lesson phases with planning, monitoring, regulation, affective engagement, embodied cognition, and evaluation.