Sheet metal forming processes are strongly influenced by process parameters, moreover fluctuations during the process and in material properties often lead to robustness problems. Therefore, numerical simulations as a function of design and noise parameters are usually adopted to detect ahead critical regions. To ensure efficiency and quality, the new era of industry 4.0 aims to develop systems that can control critical issues online. The aim of this work is developing a methodology that allows to adjust the force on the blank holder to obtain an optimal draw-in profile. This latter assures a minimization of defects (e.g., compression areas, thickened areas, insufficient stretch areas, splits) on the final component, even when variability of friction coefficient and yield strength occur. This methodology is applied to the deep-drawing process of a component obtained from a DC05 steel sheet with a thickness of 0.75 mm. For this purpose, Finite Element (FE) simulations in AutoForm software have been performed as a function of the design parameter i.e., Blank Holder Force and of noise parameters i.e., friction coefficient and yield strength. The results explored to evaluate the quality of the stamped part are: (i) percentage of compression areas, (ii) thickened areas, (iii) insufficient stretch areas, (iv) excessive thinning areas and (v) areas with splits. The data collected from FE simulations were adopted to perform a multi-objective optimization by means of the desirability function approach. Consequently, the value of process parameters that minimize defects (high value of total desirability) were identified. In correspondence with this optimal solution, the optimal draw-in profile was obtained. Such optimal profile was compared with the draw-in profile which involves defects on the final component (non-optimal draw-in profile). This latter was obtained for a value of the force on the blank holder equal to the optimal one and for non-optimized values noise parameters. In fact, noise parameters are not controlled process parameter. Based on the point-by-point difference between the two draw-in profiles, a numerical control strategy was implemented using AutoForm and MATLAB software. This iterative strategy allows to modify step by step the force on the blank holder so that the non-optimal draw-in profile match with the optimal one. Finally, experimental tests on a 3000 kN hydraulic press were performed to verify FE results.