<![CDATA[This study aims to analyze students' metacognition ability in solving PISA-based math problems on space and shape content consisting of three essay questions. Metacognition was analyzed through three indicators of metacognitive regulation, namely planning, monitoring, and evaluation. This study uses a descriptive qualitative approach with junior high school students selected through purposive techniques which are grouped into three ability categories: high, medium, and low. Data was obtained through analysis of student answers and in-depth interview results. The results of the study show that high-ability students are able to activate all aspects of metacognition optimally, characterized by the ability to formulate completion strategies, monitor thought processes, and evaluate results reflectively. Moderately capable students show mastery of metacognition at a sufficient level, but tend to be inconsistent in monitoring and evaluating the strategies used. Meanwhile, low-skilled students exhibit limited metacognition skills, characterized by minimal planning, weak monitoring, and procedural evaluations without conceptual understanding. These findings confirm that metacognition ability is directly proportional to the quality of mathematical problem solving, particularly in PISA problems that require geometric visualization and spatial reasoning. This study recommends the need for learning that explicitly trains metacognitive strategies to improve students' ability to solve context-based problems.]]>
