The suspension system plays a crucial role in ensuring the smoothness, stability, and safety of high-speed trains. This paper provides a comprehensive overview of the development of passive, semi-active, and active suspension systems. Advanced intelligent control architectures, such as Proportional Integral Derivative (PID), Linear Quadratic Regulator/Gaussian (LQR/LQG), and Sliding Mode Control (SMC), are compared with modern approaches, specifically fuzzy logic control frameworks based on Particle Swarm Optimization (PSO), Adaptive Neural Networks (ANN), and Adaptive Nonlinear Control (ANC). The paper evaluates the advantages and disadvantages of each strategy by considering five core criteria: passenger comfort, vibration suppression capability, disturbance rejection, adaptability, and implementation complexity. The synthesis of results indicates that intelligent and adaptive controllers provide significant quantitative enhancements; for instance, a PSO-optimized hybrid Fuzzy-PID controller achieves a 42.8% reduction in root-mean-square (RMS) acceleration. Most notably, the ANC strategy attains the highest improvement, enhancing ride comfort by up to 68.2% compared with passive systems. Finally, future research directions are outlined, emphasizing the necessity of high-fidelity multi-physics modeling and the development of computationally optimized, data-efficient reinforcement learning frameworks directly integrated into fault-tolerant control loops.