Latent heat thermal energy storage technologies using Phase Change Materials (PCMs) are a promising solution to enhance building thermal performance and reduce energy consumption. However, conducting experimental or numerical studies becomes time-consuming and computationally expensive due to the non-linearity and complexity associated with climatic conditions and variations in the thermophysical properties of PCMs within the building envelope. Thus, effective energy demand forecasting is essential for optimizing planning and minimizing energy consumption in buildings. Machine learning techniques have become increasingly popular due to their reliability and cost-effectiveness. This study uses machine learning models to predict the energy consumption of PCM-integrated residential building envelopes. Five well-known machine learning models were investigated: Multiple Linear Regression (MLR), Support Vector Regression (SVR), Artificial Neural Network (ANN), Generalized Additive Model (GAM), and Decision Tree (DT). These models considered PCM thermophysical properties, location, and thickness variations. The dataset was generated through a parametric analysis using EnergyPlus and JEplus simulation tools. The prediction computations were conducted using a computer program written in Python-based software. Thereafter, model performance was assessed using three metrics: R2, MAE, and RMSE. The results indicate that the ANN model outperformed others, achieving the lowest RMSE, MAE, and the highest R-squared value exceeding 0.99. Moreover, this study's findings emphasize the potential of the ANN model in predicting energy consumption and offer valuable insights for stakeholders aiming to optimize heating and cooling energy consumption in PCM-incorporated residential buildings.