Abstract: Tribological behaviors are characterized by dependence on the system, temporal evolution and multidisciplinary coupling, leading to a large number of multi-modal and multi-source research data. Researchers in different disciplines tend to understand and address tribological issues from their unique perspectives, making the tribological information entropy increasing continuously. As a result, the data in tribology are scattered and difficult to apply across disciplines. To achieve effective integration of multidisciplinary knowledge, tribo-informatics has been proposed and developed in recent years. From the perspective of tribo-informatics, the implementation path is designed for tribology and the informatics methods are utilized to address the complex information of the tribo-system. In this paper, the connotation of tribo-informatics and the data-model-fused design concept of tribo-system were elaborated. The concept outlined the dynamic evolution of tribological data, information and knowledge, as well as the design process involved in interdisciplinary integration. Based on the three tribological axioms, a universal informational expression model for the tribo-system was constructed: TriboInfo= I, S, T_s, D_s, O . In this model, five categories of information were identified—input information (I), system intrinsic information (S), tribological state information (T_s), derived state information (D_s), and output information (O). These categories were described in detail and exemplified. The essence of tribo-informatics lied in employing artificial intelligence (AI) techniques to establish relationships between different categories of tribo-system information. The main methods employed in this process include dimensionality reduction, clustering, classification and regression, which were particularly relevant in addressing four critical practical application needs within tribological field: state monitoring, behavior prediction, system optimization and mechanism analysis. In light of those application requirements, the processes and steps involved in the collection, integration, processing, application, storage, and sharing of tribo-system information were systematically outlined. This outline was based on the integration of typical models and methods of AI. A complete research framework for tribo-informatics was developed, serving as an information-based perspective and practical reference for tribological research. Through the illustration of concepts and the construction of the research framework, this study was expected to provide insights into how to incorporate informatics technology into tribological studies. It provided a more structured and systematic approach to addressing the complexities of tribological research compared to traditional methods, ultimately resulting in improved outcomes across various applications.