In order to study and optimize the platform orienting sign system of high-speed railway station, particle swarm optimization (PSO) is introduced to build the platform model. A new "air sign" system is added to the original high-speed railway landmark system. The platform simulates passengers to find the corresponding location. After adding the new sign system, PSO is used to verify that the model has higher fitness and convergence speed, and its performance is better than that of the landmark only guidance system. The optimization of platform guidance sign will use esp8266 wireless module for information transmission and display screen to display information. The new "air sign" system can be combined with dispatching information to display the current platform car number information in real time. In the training of PSO, the iterative time of the optimized guidance identification system is reduced by 44.8%, and the effect is more significant.