Cross-line trains are highly sensitive to disturbances in real-world railway operations, where small delays can easily propagate across lines and stations. To address this issue, this study proposes a hybrid intelligent optimization method that integrates Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), termed GAPSO, aiming to minimize the total delay and rapidly generate executable timetables that satisfy both temporal and resource constraints. The approach encodes all trains’ arrival and departure times as a unified time vector, incorporating constraint-aware decoding, feasibility repair, and adaptive penalty-based fitness. GAPSO alternates between PSO-based global search (odd generations) and GA-based local refinement (even generations). In real operational cases, GAPSO reduces total delay from 1049.01 to 215.43 min—far outperforming the GA-only baseline (547.86 min)—showing remarkable improvements in rhythm stability and conflict mitigation. The proposed method demonstrates strong engineering practicality and scheduling stability, providing a feasible path for “rapid rescheduling under delay scenarios” in real-time railway operations.