Building energy consumption prediction is a crucial component of energy efficiency goals. Engineering, AI-based, and hybrid techniques may all be used to anticipate how much energy a building will need, we choose the AI-based technique because it makes predictions about future energy usage within limits using historical data instead of thermodynamic equations which are used by the other approaches. Therefore, the goal of this research is to implement and evaluate some forecast models for energy use, the suggested algorithms are multiple linear regression, linear regression, random forest, and long short-term memory. The dataset of our work was collected from a commercial building that serves as a case study, and we used RMSE, MSE, MAE, and MAPE measures to compare the effectiveness of each approach.