Researchers have shown that teaching becomes very effective if learners are taught to construct knowledge rather than transform information, to monitor their own learning and progress, and to apply their learning and understanding authentically. It is therefore essential that the major implication of constructivism and metacognition should be reflected in classroom practices. This paper aimed to investigate the effectiveness of the constructivist learning model with a metacognitive strategy in teaching mathematics. A pretest-posttest quasi-experimental design was utilized incorporating both qualitative and quantitative techniques. Two intact-classes of Math 2 were the respondents of the study. The constructivist learning model with a metacognitive strategy has been applied to the experimental group while traditional methods of teaching were followed by the control group. An adapted test in Probability Theory was used as an instrument for both pretest and posttest to evaluate the student's performance. The hypothesis was tested at a= 0.05 using ANOVA. Findings revealed that the respondents were almost the same in terms of their mathematical ability based on their average grade in Math 1. Pretest mean achievements of the control and experimental groups have no significant difference. Meaning, the two groups were of equal competence before the treatment started. It was also found out that the two groups' mean achievements in the posttest have a high statistical difference favouring the experimental group. Hence, adopting a constructivist learning model with a metacognitive strategy significantly improved the students' achievements in the course as compared to using traditional teaching methods. Based on the above findings, it is concluded that the constructivist learning model with a metacognitive strategy has a positive impact on the mathematical teaching-learning process.