This paper presents a new experimental system for moving train parameter identification on bridge. The numerical method is adopted to simulate the experimental system and to investigate the effect of the system on identification of the moving train parameters. Displacement responses are obtained by solving the equation of motion of a simply supported bridge model. Basic parameters of the train (i.e., moving speed, wheel axle numbers, and axle spacing) are recognized with a self-designed parameter recognition system. Axle load values on bridge are identified with a load identification system based on simulated annealing genetic algorithm and by minimizing the errors between measured and reconstructed displacements from the moving train loads at each time step. Measured data are simulated by adding artificial noise to the calculated data. A series of numerical simulations is performed to analyze the effect of the system under various train situations. The results show that the experimental system is feasible and valid for identification of moving train parameters, thereby providing effective preparation for practice.