As artificial intelligence (AI) systems become increasingly integrated into modern applications, their environmental impact, particularly energy consumption, has raised concerns. While much research has focused on optimizing machine learning models for energy efficiency, less attention has been given to other Al areas, such as automated planning. This thesis focuses on analyzing the energy consumption of Hierarchical Task Network (HTN) planners, a prominent planning technique widely applied in real-world applications such as smart buildings, autonomous driving, and cloud computing. A systematic methodology was adopted, including the identification of distinct operational phases such as parsing, grounding, task decomposition, solving, and plan verification for each HTN planner. To measure energy usage accurately, tools like Intel RAPL and custom profiling scripts were used, enabling the analysis of energy consumption across a diverse set of benchmark domains. The planners evaluated include PANDApro, PandaDealer, TOAD, Aries, and HyperTensioN. Results reveal that the solving phase consistently consumes the most energy, followed by the grounding phase. Additionally, significant variations in energy consumption were observed between different planners and domains. This research provides valuable insights into optimizing energy efficiency in Al planning, highlighting the most energy-consuming phases where improvements can be made. These findings contribute to the broader effort of making Al technologies more sustainable, ensuring that advancements in Al align with environmental responsibility.