Predictive PI controllers have been widely used for industrial process plants with deadtime. This is due to its advantages of simple design, easy tuning, and implementation. However, this smith predictor based structure failed to achieve adequate performance in the presence of high disturbances and variation in set-point. Therefore, in this research, fractional-order predictive PI controller has been proposed for deadtime processes. The controller is based on the predictive nature of Smith Predictor and robustness nature of fractional-order controller. The proposed approach consistently manages to give a better closed-loop performance for both servo and regulatory process operations. Various real-time industrial process plant models are simulated with longer dead-time to evaluate the controller flexibility, effectiveness of handling disturbance, and time-delay compensation capabilities. Tuning rules for the proposed controller has derived analytically by using first-order plus deadtime (FOPDT) process model in frequency domain representation. © 2020 IEEE.