<p>Effective optimization methods are required due to the increasing complexity and difficulty of real-world optimization problems. Although several metaheuristic methods have been proposed thus far, only a few have gained widespread acceptance in the scientific community. To address optimization problems, this study presents a novel metaheuristic method known as the Phong optimization algorithm (POA). The POA algorithm was developed using the Phong reflection model, an intriguing model in the field of computer graphics. The Phong model is an illumination model used to produce realistic 3D images by simulating the interaction of light with surfaces. Lighting effects at specific locations on a surface are calculated using this local lighting model, which ignores overall lighting effects. Phong model is fundamental to creating realistic shading and highlights on 3D objects. These inspirations were mathematically formulated to focus on exploring and exploiting light simulation and interaction with surfaces inside a certain search space. The ability of the POA algorithm to precisely search the entire search space with good convergence speed is very significant. Three benchmark test suites, namely CEC2019, CEC2020, and CEC2022, were used to fully evaluate the performance of the proposed POA algorithm. These test sets consist of composition, multi-modal, unimodal, and hybrid test functions with different degrees of complexity and dimensionality. Several classification problems were also solved using the POA algorithm to demonstrate its dependability and suitability in practical settings. The results of the proposed POA algorithm surpassed a set of well-known state-of-the-art and newly released metaheuristic algorithms. The experimental findings of the proposed POA algorithm indicate that POA is an outstanding performance optimization tool in terms of the balance of exploration and exploitation and convergence speed, as is efficiently applicable to tackling complex optimization problems.</p>