In order to quantify the impact of depot capacity on the operation of rolling stocks in urban rail transit and make appropriate decisions on the capacity design of depots in the planning stage, this paper investigates the depot capacity design in the conditions of multiple depots, service lines, and turn-back stations under a single line. First, a mixed-integer linear programming model was established for different operation phases of rolling stocks considering departure modes, with the deadhead mileage as the optimization objective. Then, the proposed model was validated by a case study of Guangzhou Metro Line 13. Finally, the effects of arrival and departure directions, departure capacity, and turn-back operations on the deadhead mileage and capacity design under different departure modes were discussed. Results showed that the cost-saving design of depot capacity could be quickly obtained by the proposed model. Additionally, the arrival and departure directions as well as the departure capacity had impact on the capacity design, and the extent of the impact varied depending on departure modes. While the turn-back operations affected the capacity design only in the case of uniform departure mode. The results in this paper provide a new method for the design of depot capacity, which can be used to support the decision of urban rail transit depot planning in practice.