The industrial revolution has increased the use of home appliances, leading to a significant rise in household energy consumption over time. Various factors, including the number of appliances and external conditions, influence energy use. Understanding how to improve energy efficiency is crucial. Analyzing energy consumption data poses challenges due to its volume, frequency, and complexity. This study employs functional data analysis (FDA) techniques, which treat the entire data course as a single curve rather than considering each observation independently. In this paper, we present functional regression, an approach that handles functional data by extending standard regression, to estimate the relationship between residential energy usage and environmental factors. Our dataset includes real-time energy usage recorded at 10-min intervals from a monitored household. We compare the performance of functional regression models to several standard machine learning models. Empirical results show that functional regression consistently outperforms linear regression, support vector machines, and random forests, achieving the lowest RMSE and MAPE for 1-day, 1-week, and 1-month forecasts, respectively. These findings highlight the reliability of functional regression in accurately modeling dynamic energy consumption patterns.