Human walking is a fundamental activity, and understanding its intricacies is crucial for diagnosing and treating various gait abnormalities and musculoskeletal disorders. This chapter explores the classification of gait phases using advanced machine learning techniques, dividing these phases into binary and three subphases. The study draws on data from 100 individuals sourced from an open-source platform and applies two distinct training methodologies. The first method employs stratified random sampling, with 80% of the data from each subphase used for training and 20% for testing. The second method involves training on data from 80% of participants and testing on the remaining 20%. Before applying various machine learning algorithms, the dataset was evaluated using different scaling and preprocessing methods, including Min-Max Scaling (MMS), Standard Scaling (SS), and Principal Component Analysis (PCA). After proper scaling, we implemented and assessed the performance of several machine learning models, such as k-Nearest Neighbors (k-NN), Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), Naive Bayes (Gaussian, Bernoulli, Multinomial) (NB), Linear Discriminant Analysis (LDA), and Quadratic Discriminant Analysis (QDA). Each model was evaluated based on multiple metrics, including cross-validation score, Mean Squared Error (MSE), Root Mean Squared Error (RMSE), accuracy, and \(R^2\) score. In binary classification, RF achieved the highest accuracy of 97.33% and \(R^2\) score of 88.91%, with k-NN closely following with 96.98% accuracy and a cross-validation score of 0.9409. For ternary classification, RF showed the best results with an accuracy of 96.58% and MSE of 0.0743, while k-NN reached 96.04% accuracy.