Abstract: Comprehensive passenger transport hub is a high density passenger flow distribution area. Passenger flow parameters predict method is necessary to forewarn passenger flow congestions and retentions of bottlenecks. In this paper, an algorithm is proposed to predict passenger flow parameter in comprehensive passenger transport hub based on radial basis function neural network. A method based on RBF neural network is studied to achieve short-term prediction of passenger flow parameter in bottleneck. Computational experiments on the actual passenger flow data from a specific bottleneck position in comprehensive passenger transport hub showed that the proposed approach is effective to predict passenger flow parameters of bottleneck position with high forecasting precisions.