High Performance Computing (HPC) technologies are essential for a wide range of scientific research and engineering tasks. However, with the increasing scale of computation, energy consumption has become a key constraint on system efficiency and sustainability. Traditional energy-efficiency optimization methods usually focus on the selection and scheduling of hardware throttles, often ignoring the important role of task scheduling strategies in improving energy efficiency. In the last few years, with the popularization of heterogeneous computing platforms, finding an effective balance between energy efficiency and performance in task scheduling has emerged as a prominent research focus. This study introduces a task scheduling optimization strategy for high-performance computing environments, which is grounded in energy preallocation. This strategy combines the Dynamic Voltage and Frequency Scaling (DVFS) technique, leading to a reduction in both scheduling time and system energy consumption. Experimental results demonstrate that the scheduling strategy proposed in this study effectively reduces energy consumption and significantly shortens scheduling time.