Abstract: Based on the train-track coupling dynamics and probability density evolution theory， a random vibration model of train-track system considering both the crosswind and track irregularity is established. The N-dimensional hypercube point set is used to generate the discrete random frequency and phase representative point set. Based on the random harmonic function method and the harmonic superposition method， the random track irregularity excitation samples and the wind speed time-history samples are simulated respectively. The obtained two random excitations are introduced into the train-track coupling system to obtain the representative responses. The probability density evolution equation of representative response is solved using the bilateral difference method to obtain the time-varying probability density evolution distribution of the vibration response. The Monte Carlo method （MCM） is employed to verify the computational efficiency and accuracy of the proposed model. The results indicate that the probability density evolution method （PDEM） is appliable to the random analysis of vehicle-track system under double random excitations including crosswind and track irregularity. When the number of representative sample points is 300 and the screening radius is 17.7， a good calculation results can be obtained by the proposed model in this paper. The random response analysis regarding a single random input cannot express an accurate result of the vehicle-track system’s random dynamic characteristic， which indicates the required consideration of both random excitations of crosswind and track irregularity in the random vibration analysis of vehicle-track system.