The quasi-experimental, two-group (treatment and comparison), pretest--posttest study of metacognition, epistemological beliefs, and science content achievement explored the effectiveness of argument-based inquiry (ABI) instruction as opposed to traditional teacher-directed lectures and structured activities. Participants included 60 eight-grade students attending two intact classes of a middle school located in an urban area. One class was treated as the treatment group and the other as the comparison group. The students in the treatment group were taught for 13 weeks with the ABI approach while those in the comparison group were taught using traditional instruction. Data were gathered by administering the Epistemological Belief Questionnaire, Metacognitive Awareness Inventory, and Science Content Achievement Test as pretest and posttest and were analyzed using t-test and MANOVA. The findings revealed that the treatment group significantly outperformed the comparison group in terms of science content achievement, metacognition, and epistemological beliefs. Treatment group students' scores in science content achievement, epistemological beliefs (i.e., justification and development), and most of the dimensions of metacognition (i.e., declarative knowledge, planning, information management, debugging, monitoring, and evaluation) significantly improved as ABI was applied. However, the ABI approach was not found to improve students' epistemological beliefs concerning the source and tentative nature of scientific knowledge, or their metacognition with respect to procedural knowledge and conditional knowledge.