A self-calibrating total-mass controller for the neuromuscular blockade (NMB) based on the minimally parameterized parsimonious (MPP) Wiener model is presented. Using input-output data collected from the initial bolus administration until recovery, the parameters of the MPP model that are used for the controller design are chosen from a set of fifty realistic models for the effect of the muscle relaxant rocuronium in the NMB. In order to overcome modeling uncertainties a numeric self-calibration of the parameters in the MPP model nonlinearity was implemented. This step matches the anesthesiologists' mindset. The controller was extensively tested in simulation and in real cases. The good reference tracking results show the reliability of the proposed strategy and supports its use in the clinical practice.