The mismatch between the feedback current and the actual current, which widely exists in low-inductance permanent magnet synchronous motor (PMSM) driving system due to the motor's complex electrical characteristics, hinders the generation of expected current in the motor. This issue can even lead to serious out-of-control consequence threatening the surroundings. To solve this issue, conventional solution usually employs a large inductor by placing it in the power line to force the match of two currents. However, this solution not only slows down the motor's response, but also increases the system's weight, volume, and the overall cost. In this article, a novel feedback current self-correction method is proposed, aiming to correct the feedback current without relying on extra hardware. To this end, first, a novel distributed capacitance contained extended model (DCCEM) is proposed to precisely describe the motor's electrical characteristics. Then, the proposed mean value oriented current estimator (MVOCE) is established to correct the feedback current. Experimental results indicate that the current error reduced by 60–99% under different working conditions. Besides, as a particularly important case, the out-of-control consequence is successfully prevented.