Abstract: The humidity control problem in the tabacco production is discussed in this article. According to the different requirements at the individual stages of tabacco drying process, a piecewise control strategy consisting of the on-off control, tracking control, and predictive control is suggested. In view of the nonlinearity, uncertainty and the large time-delay of the tabacco drying process, a self-tuning predictive control algorithm based on the fuzzy model is applied to the stable stage of the tabacco drying process incorporating the feedforward control method. The feasibility is shown by the pertinent simulations.