<p>Structural Equation Modelling (SEM) has become a widely used analytical technique in the social sciences for examining complex relationships among latent constructs. Within SEM approaches, Partial Least Squares Structural Equation Modelling (PLS-SEM) has gained increasing attention in financial literacy research due to its predictive orientation, flexibility in handling complex models, and ability to accommodate non-normal data. This study provides a systematic and critical evaluation of the use of PLS-SEM in financial literacy research. Using a structured review protocol based on PRISMA guidelines, relevant articles were identified from the Scopus database and screened using predefined inclusion criteria. A total of 190 empirical models employing PLS-SEM were analysed to examine sampling characteristics, measurement model specifications, structural model evaluation, predictive assessment practices, and theoretical foundations. The results highlight several methodological shortcomings, including inconsistent reporting of measurement and predictive assessment procedures, limited application of predictive relevance measures such as Q² and PLSpredict, and an over-reliance on specific software platforms such as SmartPLS. The review identifies important gaps in the application of PLS-SEM and suggests future directions for improving methodological transparency and theoretical integration. In particular, future studies should incorporate higher-order constructs, adopt advanced predictive assessment techniques, and explore diverse theoretical perspectives to better capture the multidimensional and contextual nature of financial literacy.</p>