To transport passengers as soon as possible, a coordinated adjustment method for preceding trains of the delayed train was developed for urban rail transit system. Based on the interaction between train capacity and passenger demand, operational constraints and the adjustment time constraints, an integer programming model was established to minimize total passenger travel time with considering the in-vehicle and on -platform passengers. A combined dynamic programming algorithm of time selection of multiple trains at multiple stations was designed to solve this model. The numerical results show that the proposed model and algorithm are effective and the method yields the minimal passenger travel time than the other methods.