With the increasing demands for reliability, safety, and intelligence in aviation equipment, the functions and internal structural relationships of avionics products have become increasingly complex, making it challenging to determine and identify fault propagation relationships. Fault detection, identification, and localization have become significantly more difficult. This paper focuses on avionics products, constructs a data-driven fault diagnosis framework for electronic products, and proposes a Clustering-based Local Outlier Factor (CBLOF) fault diagnosis algorithm. This method achieves accurate anomaly detection while substantially reducing model complexity. Finally, validation experiments on a typical electronic module demonstrate its effectiveness.