By extending the system's state variables, a novel predictive functional controller has been developed. The structure of this controller is similar to that of classical proportional integral (PI) optimal controller and includes a control block that can perform a feed-forward control of future P-step set points. It considers both the state variables and the output errors in its cost function, which results in enhanced control performance compared with traditional state space predictive functional control (TSSPFC) methods that consider only the predictive output errors. The predictive functional controller (PFC) has been compared with TSSPFC in terms of tracking ability, disturbance rejection, and also based on its application to heavy oil coking equipment. The results obtained show the effectiveness of the controller.