The development and application of a constrained Single Input Single Output (SISO) version of the popular Generalised Predictive Control (GPC) algorithm, which uses the Quadratic programming (QP) approach, is presented in this paper; Mean Arterial pressure (MAP) is used as an inferential variable to indicate the level of unconsciousness. First, the algorithm was validated using a derived re-circulatory physiological model of anaesthesia via a semi-closed circuit before the closed-loop control system was transferred to the operating theatre for validation during surgical operations. Simulation and real-time experiments showed that excellent regulation of blood pressure around set-point targets can be achieved. Such regulation was later translated to being equivalent to a good maintenance of level of anaesthesia.