Vibration is an important index to evaluate the operating safety, comfort and performance of rail transit vehicles. In order to further obtain the vibration characteristics and characteristics of structures, modal analysis is generally used as an effective way. If the structure has a natural mode in a certain frequency range, its characteristics can be estimated and determined by modal analysis method, and the actual structural vibration response of the structure to external or internal excitation source excitation and transmission under the natural frequency range corresponding to the mode can be calculated on this basis. This paper takes rail vehicle as the research object, aiming at the test modal analysis, the layout of modal test sensor is further studied. By optimizing the layout position of modal test sensor of rail vehicle, the results of modal identification based on sensor test data are more accurate, convenient and effective. Based on particle swarm optimization algorithm, this paper proposes an improved scheme for the sensor configuration optimization existing in the current rail vehicle body modal analysis test test test. Based on engineering practice experience, combined with the actual test scenario of rail vehicle body modal test, and according to the characteristics of the three-dimensional sensor used in the actual test, an intelligent particle swarm algorithm based on modal confidence method is proposed. For the actual test of the rail vehicle body mode, the position of the measuring point obtained by the particle swarm optimization algorithm is adjusted according to the test ability of the laboratory and the consideration of the integrity of the vibration mode, so as to maximize the test ability.