As a critical component of marine electric propulsion systems, the propulsion motor frequently operates under sustained high-load conditions in harsh marine environments characterised by elevated humidity, temperature, and salinity. Such demanding conditions often precipitate component failures, while external disturbances including wake turbulence, wave impacts, and fluctuating air currents further destabilise motor speed and torque, causing non-stationary operational behaviour. Consequently, vibration signals become dispersed and exhibit spectral smearing, complicating the extraction of characteristic features related to motor bearing faults. Accurate estimation of instantaneous frequency is therefore essential to ensure effective tacholess order tracking under variable speed conditions. This study proposes a current-aided order tracking method for vibration signal analysis in propulsion motors, leveraging zero-crossing points within current signals for precise instantaneous frequency estimation and angular resampling. Envelope spectrum demodulation is subsequently employed to extract fault-specific spectral characteristics, significantly aiding the detection and diagnosis of mechanical faults. The effectiveness and robustness of the proposed method were validated experimentally using a small-scale marine propulsion system under varying speed conditions. Results demonstrated reliable identification of bearing outer race faults, substantially enhancing diagnostic accuracy and suggesting broader applicability and improved reliability for fault diagnostics in fluctuating marine operational contexts.As a critical component of marine electric propulsion systems, the propulsion motor frequently operates under sustained high-load conditions in harsh marine environments characterised by elevated humidity, temperature, and salinity. Such demanding conditions often precipitate component failures, while external disturbances including wake turbulence, wave impacts, and fluctuating air currents further destabilise motor speed and torque, causing non-stationary operational behaviour. Consequently, vibration signals become dispersed and exhibit spectral smearing, complicating the extraction of characteristic features related to motor bearing faults. Accurate estimation of instantaneous frequency is therefore essential to ensure effective tacholess order tracking under variable speed conditions. This study proposes a current-aided order tracking method for vibration signal analysis in propulsion motors, leveraging zero-crossing points within current signals for precise instantaneous frequency estimation and angular resampling. Envelope spectrum demodulation is subsequently employed to extract fault-specific spectral characteristics, significantly aiding the detection and diagnosis of mechanical faults. The effectiveness and robustness of the proposed method were validated experimentally using a small-scale marine propulsion system under varying speed conditions. Results demonstrated reliable identification of bearing outer race faults, substantially enhancing diagnostic accuracy and suggesting broader applicability and improved reliability for fault diagnostics in fluctuating marine operational contexts.