Early and accurate detection of distributed bearing faults is critical for preventing equipment failures and minimizing downtime in industrial environments. This study explores the optimal selection of input signal sources for high-accuracy distributed fault classification, employing wavelet transform and machine learning algorithms. We utilize a comprehensive dataset encompassing three-axis vibration, stray magnetic flux, and two-phase current signals to diagnose six distinct bearing fault conditions. Wavelet transform is applied to extract frequency-domain features, capturing intricate fault signatures. To identify the most effective input signal combinations, we systematically evaluate Random Forest, XGBoost, and Support Vector Machine (SVM) models. Our analysis reveals that specific signal pairs significantly enhance classification accuracy. Notably, combining vibration signals and stray magnetic flux consistently achieves the highest performance across models, with Random Forest reaching perfect 100% test accuracy and SVM demonstrating robust performance. These findings underscore the importance of optimal source selection and the effectiveness of wavelet-transformed features for improved machine learning model performance in fault classification tasks, contributing to more reliable and cost-effective predictive maintenance systems for industrial machinery.