<p>This study aims to develop a model-free type intelligent backstepping control-based polynomial Petri fuzzy neural network (IBC-PPFNN) in order to construct an optimal synchronous reluctance motor (SyncRM) servo drive system because of the unexplained nonlinear dynamic nature of the SyncRM. First, an ANSYS Maxwell-two-dimensional control of a SyncRM servo drive with maximum torque per ampere (MTPA) is presented. To generate the direct-axis current command of the MTPA, a lookup-table (LUT) is constructed via the results of the finite element analysis (FEA). Then, position references are tracked by a nonlinear backstepping control (NBC) system. However, it is difficult to create a workable NBC for real-world applications because of the complex system dynamics and the unexplained nature of the SyncRM servo drive system, which is not disclosed in advance. To achieve a model-free approach, a polynomial Petri fuzzy neural network (PPFNN) is proposed as an alternative to the ideal NBC. Furthermore, an adaptive compensator is developed in order to immediately handle the estimated errors of the PPFNN. To guarantee asymptotic stability, the online learning algorithms are obtained via the Lyapunov stability proof for the PPFNN. Finally, the study’s conclusion includes the results of experiments proving the practicality and efficiency of the proposed IBC-PPFNN controlled SyncRM servo drive.</p>