Permanent magnet synchronous motors (PMSMs) are crucial components in power systems, directly influencing their safety and reliability. Stator winding faults are among the most common and serious issues affecting PMSM operation. Due to the complexity of operating conditions and significant fluctuations in fault characteristics, traditional diagnostic methods often suffer from low accuracy. This chapter proposes an online stator winding fault diagnosis method based on machine vision. By analyzing the mechanisms of typical winding faults and collecting data on winding voltage differences and currents, a fault model based on ΔU–I Lissajous figures is developed. Research results demonstrate that ΔU–I Lissajous figures effectively distinguish winding faults, and the machine vision-based approach enables rapid assessment of fault severity with high accuracy.