Based on the multivariable description of processes in industry, a non-zero-pole cancellation decoupling strategy is first proposed and then used for state space predictive functional control (PFC) design. The proposed decoupling guarantees realization and enables the control system design to be based on single-input single-output (SISO) process formulations. To facilitate the closed-loop control performance improvement, the subsequent controller design adopts an extended non-minimal structure that can regulate the process state dynamics, which provides more degrees compared with traditional state space methods. By interpretations of the proposed performance through process transfer function formulations, relationship with and superiority to traditional state space PFC are further revealed. Finally, the effectiveness and merits of the proposed are illustrated by application to a typical industrial chamber pressure process, in comparison with a typical non-minimal state space PFC method recently developed. (c) 2013 Elsevier Ltd. All rights reserved.