As the need for improved machinery reliability grows, the role of artificial intelligence (AI) in fault diagnosis becomes increasingly important. However, there is limited research systematically comparing the performance of different machine learning algorithms for monitoring rolling element bearings. This study addresses this gap by comparing seven widely used machine learning algorithms in MATLAB: Gaussian Support Vector Machine (GSVM), Ensemble Bagged Trees (EBT), weighted K-Nearest Neighbors (WKNNs), Fine KNN (FKNN), Subspace KNN (SKNN), Wide Neural Network (WNN), and Artificial Neural Networks (ANNs). Experiments were conducted using a machinery fault simulator under various conditions, including healthy and faulty states in both the outer and inner races, at multiple speeds. Faults were induced using Spark EDM on SKF ball bearings. Vibration data from the test bearings were acquired using the National Instruments data acquisition system with LABVIEW. Statistical features such as RMS Value, kurtosis, Crest Factor, and Impulse Factor were extracted from these vibration data using MATLAB to create a robust dataset for training the machine learning algorithms. Results indicate that the ANN, WNN, and EBT models consistently achieve the highest accuracy during the validation phase. Conversely, the remaining four models exhibited comparatively lower accuracy at this stage but showed notable improvements during testing. Remarkably, the EBT, WKNN, FKNN, and SKNN models attained 100% accuracy across all bearing classes in the testing phase. These findings highlight the exceptional performance of the WKNN and ANN models, positioning them as top-performing algorithms among the seven evaluated. The study underscores the potential of these machine learning models in providing precise and reliable bearing fault classification, which is instrumental for enhancing predictive maintenance strategies in industrial applications.