Abstract: The running time of urban rail transit train air-conditioning condenser fan is quite little, basically turned-off except in summer, and the maintenance schedule may have arranged excessive repair for this condition, therefore the air-conditioning condenser fan maintenance schedule needs to be optimized. Targeting this issue, a data-driven maintenance schedule optimization strategy for condenser fan is proposed. Through this strategy, based on the effective service mileage of condenser fan components during the actual operation of urban rail transit vehicles, the fault data is obtained. A hypothesis test is carried out on the fault data to determine the distribution type suitable for the component lifespan. The two-parameter Weibull distribution is used to fit the fault data, and the support vector regression machine suitable for the fault data types of condenser fan is introduced to fit the fault data to solve the distribution parameters. Then the reliability model is constructed according to the parameters, and the reliability interval constraint is obtained when the reliability decreases. According to the labor cost required by condenser fan preventive maintenance, the labor and component maintenance costs in case of failure, the preventive maintenance and repair maintenance costs of the component are calculated separately, and the total maintenance cost model of the component is obtained. Through the analysis of comprehensive reliability model and maintenance cost model, the optimal maintenance schedule is obtained, that is, under the condition of meeting the reliability of components, the service mileage with the minimum total maintenance cost is the optimal maintenance mileage. Results show that the optimized vehicle condenser fan maintenance schedule is more reasonable than the original schedule, which can not only ensure the safe operation of the train, but also reduce the inspection and operation-maintenance cost.