Within the large Additive manufacturing (AM) process family, Directed Energy Deposition (DED) can be used to create low-cost prototypes or to repair a crack in a part. For M4 HSS (High Speed Steel), which is one of the most commonly used steel grade for tools, an important issue in DED manufacturing is that during the printing process, residual stress can lead to failure if the substrate is not properly preheated. To obtain an optimal preheating, experimental calibration would considerably increase the manufacturing cost, where numerical simulations can be useful by predicting the temperature field with a numerical model [1]. However, 3D AM simulations taking into account all the passes to follow the physical phenomena are generally unaffordable due to huge CPU time. Thus, the 2D finite element assumption can be useful (it can be more than 20 times faster) [2]. However, to get a representative 2D temperature field, the simulations must be accurately calibrated with both material thermophysical properties and numerical parameters, such as the mesh density and the boundary conditions. The present study assesses the impact of different choices on the predicted melt pool depth and cooling rate.