This paper considers control systems in the presence of actuator saturation. A traditional approach to alleviating the adverse effect of actuator saturation is to design an anti-windup mechanism and activate it as soon as the saturation occurs. A recent innovation is to delay the activation of the anti-windup mechanism until the degree of saturation reaches to a certain level. By allowing the nominal controller to function undisturbed in the face of modest actuator saturation, delayed activation of the anti-windup mechanism has been shown to lead to improvement in the closed-loop performance. In this paper, we propose to activate the anti-windup compensation in anticipation of actuator saturation and show that such a “precautionary” approach has the potential of leading to further improvement in the closed-loop performance, in terms of the transience quality in reference tracking and the region of stability.