Traditional fault diagnosis methods, such as Time-Domain Reflectometry and Frequency-Domain Reflectometry, often struggle to handle complex fault signals and have limitations in accuracy and real-time performance. This research aims to develop a more effective cable fault diagnosis model that combines wavelet transform and fuzzy reasoning to improve detection accuracy and real-time performance. The proposed model uses wavelet transform for multi-scale decomposition of fault signals, extracting high-frequency and low-frequency features, while the fuzzy reasoning system classifies and diagnoses the fault signals based on a preset rule base. Experimental results show that the model achieves high accuracy in identifying various fault types, including short circuit, grounding, open circuit, and partial discharge, with a short circuit fault accuracy of 94.5% and an average diagnosis time of 0.8 seconds. The model also demonstrates strong robustness under noise interference, maintaining over 90% classification accuracy even at a noise intensity of 20 dB. Compared to traditional methods, the model excels in handling complex faults and multiple signals while maintaining high noise resistance. Future research will focus on enhancing real-time performance, improving rule base design, and expanding the model’s ability to handle multi-fault scenarios.