Maintainability and serviceability are the most important factors which influence the delivery of Product Service Systems (PSS). To enhance the maintainability and serviceability for collaborative service in the context of PSS, a decision-making method for the delivery management of PSS is urgent needed both in practical and theoretical aspects. This study proposed a method to solve a collaborative service decision-making problem for the delivery management of PSS and to assess its performance. Several service stations with specific service area are supposed to satisfy the collection of service requirements from several customers in the joint area collaboratively by sharing vehicle fleets, service workers, and so on. A mathematical model was built with constraints of technician number, skill, time windows, maximum driving distance for the vehicles, and collaborative service options. A hybrid algorithm of PSO (Particle Swarm Optimization) and GA (Genetic Algorithm) is given to solve this practical large-scale problem. An empirical study of agricultural machinery maintenance clearly demonstrated that the proposed methods are effective and the algorithms can provide reasonable solutions within an acceptable computational time. Moreover, by comparing the total cost of the independent service mode with the collaborative mode, we concluded that the collaborative service total cost outperforms the non-cooperative one. Undoubtedly, our method is quite effective and can be used to improve the performance of collaborative service decision-making for the delivery management of PSS.