Passenger comfort in railway vehicles is strongly influenced by the dynamic behavior of the suspension system. This study investigates the optimization of PID-controlled semi-active suspensions using metaheuristic algorithms, including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Water Flow Optimization (WFO), and Starfish optimization algorithm (SFOA). A quarter-vehicle model (QVM) with primary and secondary suspensions were developed using MATLAB/Simulink. The system dynamics were modeled using Newton’s second law, and results were validated through Simscape Multibody simulations. The performance of the semi-active suspension system was evaluated based on the root mean square (RMS) of vertical accelerations and rail vertical offsets. Metaheuristic algorithms were applied to optimize PID gains, aiming to minimize RMS accelerations while satisfying system constraints. Results demonstrate that all algorithms significantly improve ride comfort compared to passive suspensions, with SFOA achieving the lowest RMS accelerations and best suppression of vertical offsets. This study highlights the effectiveness of metaheuristic-based PID optimization in enhancing passenger comfort and provides guidance for the design and implementation of advanced semi-active suspension systems in railway vehicles.