This paper describes a practical method for moisture control in fluidized bed granulation by means of a neural network. Wet granulation of pharmaceutical powder was conducted using an agitation fluidized bed, and moisture content was continuously measured by IR (infrared) moisture sensor. A neural network system for moisture control was developed using moisture content and its changing rate as input variables, and the moisture control characteristics were investigated by the neural network system with a back propagation learning. Good response and stability without overshoot were achieved by adopting the developed systems. This system also maintained favorable stability under various operating conditions.