The purpose of this study was to investigate the metacognitive functioning of students with learning disabilities (LD), low-achieving (LA) students, and average-achieving (AA) students within the context of math problem solving. Metacognition, that is, the awareness individuals have regarding their own mental processes and ability to self-regulate performance, is an important predictor of learning. Deficits in metacognition have been attributed to an inability to effectively balance the cognitive and metacognitive strategies necessary for successful problem solving. Students with LD have considerable difficulty with self-regulation. This study investigates three components of metacognition: metacognitive knowledge, metacognitive experience, and metacognitive skills. The differences in these components among students with LD ( n = 15), LA students (n = 38), and AA students ( n = 29) and their influence on students' math word problem solving was studied. Furthermore, the relationships among the three components of metacognition were investigated in the context of ability group differences. To assess metacognitive functioning, students were administered a structured interview and a survey and they solved three math word problems while thinking aloud. Additionally, to assess math problem-solving ability, students were administered a 10-item math word problem-solving test. Results indicated that students with LD demonstrated a different pattern of metacognitive function than AA students and LA students. Students across ability groups look relatively equivalent in the quantity of metacognitive skills. However, when discriminating between the type and quality of the metacognitive skills employed, ability group differences were evident. Ability group differences in metacognitive functioning emerged with respect to problem difficulty. The directions of the relationships among the components of metacognition were the same across ability groups. However, the magnitude and strength of the relationships differed by ability. Additionally, metacognitive knowledge was a significant predictor of math word problem-solving performance for AA students, but not for the other ability groups. Furthermore, there was a significant difference in the relationship between metacognitive experience and math word problem solving for students with LD and AA students. Educational implications are discussed for teaching students to use metacognition during problem solving.