Abstract: As invoking and execution of manufacturing resources is more complicated than computational resources, the service combination execution reliability becomes a mandatory considering problem in cloud manufacturing.In this paper, we present a cloud manufacturing service composition method considering execution reliability based on discrete particle swarm optimization algorithm.The method described the service reliability using the probability density function, and generated workflow model of the cloud manufacturing services combination dynamically based on service model base.The matching service could be achieved according to the semantic.The combination scheme of cloud manufacturing services with optimal execution reliability could be calculated using the particle swarm optimization algorithm with discrete variables.Finally we take the structure design workflow model for example, the validity and reliability of the proposed method were demonstrated considering the different service instances size and malicious services accounted cases. The results also showed that the method could still maintain the high service execution reliability as the malicious services accounted was much higher.