This study presents a bibliometric review of portfolio optimization research from an Operations Research perspective, covering literature published between 2015 and 2026. A final corpus of 708 documents was identified through a PRISMA-based selection process using Scopus. Biblioshiny and Microsoft Excel were used for data extraction, analysis, and visualization. The study examines publication and citation trends, source performance, author productivity, institutional and country-level output, international collaboration, keyword patterns, conceptual clusters, and thematic evolution. Four research questions address publication and citation activity; prominent journals, authors, institutions, and countries; the field’s main conceptual and methodological approaches; and changes in themes between earlier and later periods. The results show that portfolio optimization and portfolio selection remain central to the literature. Robust optimization, robust portfolio optimization, genetic algorithms, asset allocation, machine learning, portfolio management, multi-objective optimization, and Conditional Value-at-Risk also appear among frequently observed concepts. Publication output is concentrated in a relatively small group of sources, while researchers and countries contribute across a broad international base. China, the United States, and India are major contributors, and the collaboration analysis shows links among leading research-producing regions. The thematic analysis indicates continuity in the field’s core portfolio optimization theme alongside increasing specialization in uncertainty modelling, advanced risk measures, metaheuristics, machine-learning applications, and dynamic stochastic models. The review identifies research opportunities at the intersection of robust or distributionally robust optimization and machine-learning-integrated decision models. It highlights the need to compare these techniques with classical mean-variance models using consistent datasets, realistic constraints, and common evaluation standards. Future research directions include out-of-sample testing, method benchmarking, transparent data definitions, ESG integration, emerging computational approaches, and broader geographic collaboration. Findings remain bounded by the Scopus-based search strategy and the 2015–2026 observation window; citation-age effects and the partial nature