<p>Service composition and optimization selection (SCOS) as an integral component of cloud manufacturing has garnered increasing research attention in recent years. However, most studies on this concept have predominantly focused on optimizing manufacturing services, largely disregarding logistics services, which play critical roles in SCOS. Furthermore, uncertain events during task execution, which can significantly impact the operational efficiency of cloud manufacturing systems, are overlooked. To address this, the current study introduces a cloud manufacturing and logistics service composition and optimization selection (MLSCOS) model. Notably, this model collaboratively optimizes both manufacturing and logistics services, emphasizing the robustness to enhance the anti-interference capability of the cloud manufacturing system. The model incorporates alternative services and considers resource reservation costs associated with these alternative services. Furthermore, it considers the probability of service disruptions and the additional logistics costs resulting from these disruptions, aiming to select alternative manufacturing services for specific manufacturing sub-tasks. To solve this model, an innovative enhanced non-dominated sorting genetic algorithm II (ENSGA-II) framework is proposed. This algorithm enhances local search capabilities through a variable neighborhood search strategy and adopts an adaptive crossover and mutation strategy to preserve high-quality solutions and avoid local optima. Benchmark function experiments and case studies demonstrate that the ENSGA-II framework exhibits good convergence and diversity, effectively addressing various MLSCOS problems.</p> Graphical abstract <p></p>