Abstract: In order to distribute the out-of-gauge freights reasonably to paths in railway network, a multi-objective optimization model was built for route selection of multi-direction out-of-gauge freights. In the model, the minimum transport route mileage, the minimum haulage time, and the minimum interference to the normal operation of the existing railway lines were taken as targets; and the distance between railway out-of-gauge freights and structure gauge, the railway transport capacity, and the loading capacity of the bridge along the way were used as constraints. According to the model characteristics, a multi-objective hybrid genetic algorithm was proposed to solve the model. In the algorithm, the natural array coding mode, together with cross and mutation operators, were designed to fit the constraints, and a simulated annealing strategy was introduced to enhance its neighborhood search capability. In addition, the proposed method was applied to the complex network containing 14 node stations and 23 sections to verify its validity. The application results show that the objective function values obtained by the proposed model and algorithm are superior to those obtained by genetic algorithm and annealing algorithm. Therefore, this method provides a technical measure for the decision-making of path selection in out-of-gauge freights transportation.