Additive manufacturing (AM) enables the repair of worn tools, providing a second life to the tool and presents real benefits in terms of economy and ecology. On the other hand, this multi-physics problem involves the coupling of thermal, mechanical, and metallurgical aspects during the repair process, making the study quite complex. In this context, both numerical and experimental approaches have been widely conducted to enhance the understanding of the multi-physical issues behind the process. In comparison to experimental measurements, the numerical method is more efficient and can provide a greater quantity of data at all points of the structure. In this study, the meso-scale thermal-fluid models and the part-scale thermal-mechanical models, based on the finite element method, are employed to investigate tool repair using the Wire-Laser Additive Manufacturing (WLAN) process. A coupling schema is proposed to improve the predictivity of numerical models. The meso-scale model is employed for temperature prediction, and then the temperature results will be used for heat source calibration of part-scale model. The part-scale model will be used to predict metallurgical proportion, hardness, deformation and residual stresses. The comparisons of experimental measurements and the numerical results show the efficiency of proposed models.