Many students still experience difficulty with academic writing as part of meeting their academic demands. The purpose of this study was to test the validity of the metacognitive writing strategy instrument and to test the effectiveness of metacognitive writing strategies in improving academic writing skills in tertiary institutions. In total, 500 students from three universities in West Java, Indonesia participated in this study. The method used in this research was factorial design analysis. Confirmatory factor analysis was used to answer the problem formulation. The data collection instruments used were the metacognitive academic writing strategy questionnaire and academic writing tests on several topics. The research findings show that, first, the questionnaire instrument using metacognitive strategies in the writing process was proven to be valid because it was able to assess the students’ metacognitive awareness. Second, metacognitive strategies make a significant contribution to the students’ academic writing abilities. Metacognitive strategies act as high level constructs and positive correlations were found between the dimensions of the strategies. The dimensions of metacognition found to correlate with and contribute to one another were the declarative, procedural, conditional, planning, monitoring, evaluation, information management, and debugging strategies. From the comparison of the proposed models, the second model (the eight-component correlation model) was determined to be better at improving academic writing ability than the first model. The second model described all components as significantly contributing more comprehensively than the first model. This research concludes with the implication that academic writing ability can be improved by increasing students’ metacognitive awareness and reflective abilities in relation to their own writing process.