<p>In the rapidly evolving field of supplier selection research, numerous optimization models and algorithms have been developed to tackle the complexities of this challenge. This paper introduces a robust optimization model based on a scenario-based mixed-integer programming approach specifically designed for supplier portfolio selection. The focus on robust optimization is motivated by the proven effectiveness of the return margin parameter in managing the inherent uncertainties associated with this problem. To effectively apply this complex model in real-world scenarios, we implement a customized hybrid metaheuristic algorithm that combines the artificial fish swarm algorithm with simulated annealing. This algorithm is rigorously tested using large-scale instances that include both uncertain and deterministic conditions in supplier portfolio selection. Our computational results benchmark the customized hybrid metaheuristic against state-of-the-art algorithms, including simulated annealing, genetic algorithms, the grey wolf optimizer, and the original fish swarm algorithm. The study also incorporates comprehensive sensitivity analyses, risk assessments, and evaluations of key parameters that impact the robust supplier portfolio selection problem. The numerical results reveal distinct risk profiles between uncertain and deterministic conditions, highlighting that uncertain scenarios present higher risks for return on investment. Despite these increased risks, the findings closely reflect real-world conditions, underscoring the practical relevance of the model. Furthermore, comparative analysis demonstrates the superior performance of the hybrid metaheuristic algorithm, acknowledging its longer computational time as a necessary trade-off for exceptional efficiency compared to other algorithms.</p>