Abstract: The control of temperature and humidity of fen-flavor Oolong-tea roaster is always cross-coupling and severely lagging, so it is difficult to obtain the accurate precision of temperature and humidity due to its uncertain model parameters. In this study, the fuzzy reasoning synthesizing rule was adopted to propose an effective cross-decoupling means and obtain more accurate results for temperature and humidity control, which can better solve the cross-coupling and the lagging problem of the control object. MATLAB simulation results showed that this method exhibits fast response, high control precision and small overshoot. Besides, it is need to establish an accurate system model so it can achieve the precise temperature and humidity control of the tea baker. Practical operation indicated that the error of the temperature and humidity can be controlled within the desired precision range, and the temperature control error was not more than 1℃, while the relative humidity error was not more than 2.5%.