Poor control of the steam generator water level in the secondary circuit of a nuclear power plant can lead to frequent reactor shutdowns. Such shutdowns are caused by violation of safety limits on the water level and are common at low operating power where the plant exhibits strong non-minimum phase characteristics and flow measurements are unreliable. There is, therefore, a need to systematically investigate the problem of controlling the water level in the steam generator in order to prevent costly reactor shutdowns. This paper presents a general framework for addressing all aspects of this problem using model predictive control techniques.