Wire arc additive manufacturing (WAAM) process is a metal additive manufacturing (AM) technology that is becoming increasingly important in Industry 4.0 as it allows for high deposition rates and the manufacture of large components. This translates in a reduction of production time, costs, and waste, while increasing material efficiency, compared to other AM technologies. Nevertheless, the WAAM process is difficult to automate, and the final characteristics of the produced part depend on the deposition parameters and intrinsic physical phenomena. Finite element (FE) modelling could represent an effective tool to tackle such issues, since it can be used to simulate the process for different combinations of deposition parameters. Nonetheless, some strategies are required to reduce the cost of simulations. This paper presents a metamodel based on basis spline entities to approximate the thermal response of the WAAM process for different combinations of deposition parameters. The metamodel is obtained as a solution of a constrained non-linear programming problem aiming at minimising the distance between the points composing the dataset (obtained via non-linear high-fidelity thermal analyses) and the fitting hyper-surface. The strategy (with the dedicated algorithm) is developed to fit with a prescribed accuracy the results of the high-fidelity non-linear FE model of the WAAM process by minimising the computational effort and cost. Lastly, a global sensitivity analysis based on Sobol’s indices is conducted using the fast-high-fidelity metamodel to investigate the influence of the main process parameters on the observed output responses, i.e., the temperature at the thermocouples placed at given locations.< Réduire