Cellular automata (CA) models are the idealization of physical systems with assumed discrete space and time and a finite number of discrete state transition of cells. Utilizing the CA model, we can capture and reveal the dynamics of railway traffic system. In this paper, the improved CA models are proposed to demonstrate the dynamic behavior of trains with conflict resolution. We propose the CA models to simulate the railway traffic network in the moving-block system, and reschedule trains by applying such two scheduling strategies as FCFS (first come first served) and FLFS (first leave first served). We analyze the relationship between the total delay and time interval of departure utilizing the proposed models. The simulation results demonstrate the rational conclusion that FLFS can cause less delay than FCFS. The proposed models can facilitate finding which plan can achieve the maximum transport efficiency for the railway network.