<p>Despite motors being essential general-purpose machinery, current methods in motor equipment health management fail to provide timely and effective responses to abnormal conditions. Case-based reasoning (CBR) can address sudden failures swiftly through structured knowledge bases, yet it faces challenges such as difficulties in feature selection and inaccurate similarity measurements, especially with class-imbalanced data. This study proposes a novel feature attribute reduction module using an improved fuzzy association rule mining (FARM) in the CBR framework. By introducing fuzzy theory to handle sharp boundaries in datasets and refining the traditional probability-based association rule mining, the F-Apriori algorithm is established to effectively identify the correlation between feature attributes and failure types in class-imbalanced datasets, enabling the extraction of strongly associated features and the optimal attribute sets. Finally, comparative experiments demonstrate that the proposed model achieves lower computational time costs and higher accuracy. Specifically, the computational time of the proposed model is reduced by 231.4 ms and 168 ms, respectively, compared to the baseline methods. The model attains an average accuracy of 90.2%, outperforming the counterparts by 4% and 2.6%, respectively. This approach effectively enhances current health management technologies and significantly improves the operational management of mechanical equipment in industrial settings.</p>