Direct field orientation principle based upon estimation of either the rotor or stator flux vector from the motor terminal voltage and current is approach that is very attractive for high motor speed range. However, at low speeds the voltage drop across stator winding resistance is higher than the estimated motor back–emf and becomes significantly causing the accuracy of the estimate flux vector to be sensitivity to the estimated stator winding resistance. This research proposed an adaptive stator flux vector observer which combined a programmable LPF with quasi–fuzzy stator winding resistance estimator to increase the accuracy for stator flux vector estimation, especially in low speed range. The simulation results show that the advantage of this driving system is adaptation of stator winding resistance parameter according to both speed and stator current of motor and ambient temperature for fast response, good transient performance, insensitive to variations in external motor temperature disturbance. In addition, the result of the simulation shows that the adaptive flux vector observer has provided the no oscillation and trace of estimated stator winding resistance compared with more accurate stator winding resistance. Moreover, the motor speed waveform does not oscillate at very low speed range both during transients and in steady state. The experimental results from prototype have shown that the system works as designed.