Abstract: Through analyzing the long-distance transportation modes of refined oil products, and considering the combustibility of refined oil products, a multi-objective programming model is established for selecting a proper transportation mode with low cost and risks. The model is solved by particle swarm optimization. When the model is applied to a refined oil products transportation company in Beijing, it is found that railway transportation spends longer time than road transportation when there is a small gap between weights of transportation cost and safety. Road transportation is preferred when the distance is short in view of time value, while less-risky railway transportation is recommended when the distance is long. The calculation results for this example show that railway transportation is better when the distance is beyond 500 km.