<p>The widespread use of advanced automation and sensors in physical factory together with the Information and Communication Technology (ICT) having the provisions for ubiquitous computing, leads to the development of Smart Manufacturing (SM) and Cyber Physical Production Systems (CPPS), as envisioned in Industry 4.0, to combat the rising business challenges. SM and CPPS are also equally important in the context of Industry 5.0, which focuses on holistic development of the society. It is undeniable that incomparable performance can be demonstrated, if the resources in a CPPS are capable of better problem-solving using higher order cognitive abilities, which is critical for efficient man–machine collaboration proposed in Industry 5.0. In this paper, a novel system level cognitive digital twin of machine tools (CDTMT) is proposed for the evaluation and ranking of machine tools based on four cognitive properties such as proactiveness, perturbation, awareness, and intuition. The various performance measures of these attributes, obtained from varieties of sub-attributes, are used to determine the overall cognitive ability of the machines and ranked accordingly. This ranking would help rationalize selection of machine(s) to work along with human in a CPPS. A CPPS comprising four turning centers, envisaged as Machine Tool 4.0, are considered for performance evaluation. Two unsupervised machine learning algorithms, namely Agglomerative Hierarchical Clustering (AHC), and a two-tier Principal Component Analysis (PCA), are used for evaluation and ranking, which yield comparable results. The results of sensitivity analysis reveals the robustness of the model.</p> Graphical abstract <p></p>