Metaheuristic optimization algorithms have emerged as indispensable tools for solving complex, nonlinear, and multi-constrained design problems in mechanical engineering. This paper presents a comprehensive systematic review, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, of the application of metaheuristic algorithms across six critical mechanical engineering domains: structural design and topology optimization, manufacturing process optimization, thermal and energy systems design, vibration and dynamic analysis, mechanism and robot design, and fatigue and Reliability-Based Design Optimization (RBDO). A total of 312 peer-reviewed articles published between 2000 and 2025 were identified from Scopus, Web of Science, ASME Digital Collection, IEEE Xplore, and ScienceDirect. The review systematically classifies metaheuristic algorithms into evolutionary, swarm intelligence, physics-based, and human-based categories, and benchmarks their performance on canonical mechanical engineering problems including truss optimization, CNC machining parameter selection, heat exchanger design, and vibration control. Key findings indicate that hybrid metaheuristic approaches, particularly those integrating surrogate models with population-based search, consistently outperform standalone algorithms by reducing computational cost by 40–70% while maintaining solution quality within 1–3% of known optima. The review further identifies that Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and the Jaya Algorithm (Jaya) remain the most widely adopted methods, while newer algorithms such as Grey Wolf Optimizer (GWO) and Whale Optimization Algorithm (WOA) demonstrate competitive performance in specific subdomains. A bibliometric analysis reveals exponential growth in publications since 2015, with emerging trends toward digital twin integration, physics-informed optimization, and Industry 4.0 applications. Critical challenges including computational scalability, multi-physics coupling, and the real-world validation gap are discussed, alongside future research directions encompassing AI-driven surrogate-assisted optimization and multi-scale sustainable design.