The performance of fleet in the mission depends on mission requirements and the health states of all individuals. Addressing the maintenance decision-making problem for equipment fleet executing continuous missions, an improved fleet-level multi-objective selective maintenance planning method is proposed that considers mission requirements and system reliability. The method uses a virtual age model to describe the lifespan of components and employs conditional survival probability to predict the performance of equipment in a given health state during missions. A relationship between fleet mission capability and equipment health states is established through universal generating function. The optimization model aims to minimize maintenance costs and maximize the probability that the fleet meets mission requirements. The model is optimized using the NSGA-III algorithm, and the advantages of this solution method are demonstrated through comparisons with other algorithms.