Abstract: Based on the exiting literatures, this paper proposed 5 latent variables and 16 observed variables, including interoperability, reliability, flexibility, quality management and cost control, to assess supply chain performance, combined with the characteristics of large and complex product manufacturing cluster supply chain. Then, the four possible performance evaluation models of cluster supply chain are built with structural equation modeling method, namely the order modes of the single factor, the order mode of latent variables unrelated, the order mode of latent variables related and second-order model. Because the exploratory factor analysis method cannot establish the consistency of performance evaluation of the dimensions of its system, so the confirmatory factor analysis method validation and filtering from the match and stability of the four possible evaluation modes. The results show that the fit degree between the second-order confirmatory factor analysis model and the sample data is the best, so the second-order confirmatory factor analysis model is an effective and reliable performance evaluation model of cluster supply chain.

Key words: large and complex products, cluster supply chain, performance evaluation, model