In response to the ‘dual-carbon’ goals and the vision of Industry 5.0, sustainable cloud manufacturing service composition aligns with national strategies and enterprise needs. This study proposes a sustainability-oriented service composition decision model. First, a sustainability evaluation index system is developed based on the triple bottom line principle and service lifecycle theory, and indicator weights are determined using the fuzzy best worst method. Next, aggregated sustainability values of service composition are calculated by incorporating task structures and logistics factors, forming a multi-objective optimization model. Then, a sustainability evaluation-based multi-objective artificial hummingbird algorithm (SE-MOAHA) is developed to solve the model, and a comprehensive evaluation method is employed to identify the optimal solution. Finally, a case study of a complex robotic arm task from the Alibaba cloud manufacturing platform validates the model’s effectiveness and superiority. Comparisons with the original MOAHA and the classical U-NSGA-III further show that the proposed method achieves superior solution quality, efficiency, and stability.