The estimation of measurement uncertainty is a critical challenge in metrology. In recent years, research efforts aimed at developing or improving uncertainty estimation methods have increasingly focused on rapidly evolving tools such as artificial intelligence, machine learning, and neural networks, as well as on the development of virtual machines and digital twins. Extensive databases are being developed to serve as the input and foundation for model training. What is crucial here is not only the accurate identification of relevant factors, but also their parametric representation, which enables the determination of their weights with respect to their influence on the output. At the Laboratory of Coordinate Metrology, a project aimed at developing a Machine learning-based expert system that will ultimately constitute a new tool for modelling coordinate measurement uncertainty is currently being carried out. This article presents the selection of factors influencing measurement uncertainty for both optical and tactile coordinate measuring systems, including: the measuring machine, software, object parameters such as size and shape, the measured feature, the measurement method, and the uncertainty estimation method. The methodology and preliminarily evaluated weights of individual factors are also discussed in the context of uncertainty estimation using a statistical analysis – multiple regression. The discussion is preceded by an introduction that reviews classical uncertainty evaluation methods, including the comparative and multi-position methods, as well as virtual and simulation techniques, and the software employed in the uncertainty estimation process. The concept of uncertainty estimation based on archival data and a parametric model is also outlined.