A nonlinear predictive controller is presented in the paper. The greatest advantage of predictive functional control (PFC) is very fast and easy calculation of control variable. Combined with the Takagi-Sugeno (TS) fuzzy model, the PFC becomes nonlinear and is called fuzzy PFC (FPFC). The controller was evaluated by simulation on the well-known process, the continuous stirred-tank reactor (CSTR), which exhibits strongly nonlinear characteristics. The TS model of the process was obtained by identification. At the end the resulting controller has been compared to a conventional PI controller.