Brake-by-Wire (BBW) systems have reached a significant level of maturity, both in their design and control. One of the topics still open for research is the parameter calibration of their controllers, which suffer the issue - particularly relevant in BBW actuators - of system parameter variation and uncertainty. Data-driven approaches are an interesting method to address this problem, where the controller parameters are tuned directly leveraging the collected data, and avoiding the intermediate step of building a system model. When applied to BBW systems, faster and robust parameters calibration have been successfully achieved. Following this research path, in this paper we propose a data-driven tuning approach, based on the Bayesian Optimization, for the parameters of the anti-windup scheme of a cascaded controller of an electro-hydraulic brake-by-wire system. Experimental results show that the proposed iterative tuning strategy is capable of automatically providing the controller parameters, yielding the target closed-loop performance. Additionally, a robust-oriented tuning procedure is discussed, which allows the controller to cope with the system non-linearities.