In modern wireless communication networks, the effective application of power-saving technologies is crucial for improving energy efficiency and extending the lifespan of devices. Therefore, as the number of 5G base stations implementing power-saving strategies increases, how to effectively distinguish the degree of intelligence of different energy-saving scheme becomes an important issue. This paper proposes a method for evaluating power-saving schemes based on temporal clustering analysis, using the k-medoids algorithm to cluster three power-saving modes: Carrier Shutoff, Channel shutoff, and Symbol shutoff. We employ three similarity measures: Euclidean distance, Dynamic Time Warping (DTW), and Soft Dynamic Time Warping (SoftDTW), to effectively distinguish between different power-saving schemes. Experimental results show that the proposed method significantly enhances the optimization of energy-saving strategies and the overall energy efficiency of communication systems, providing a robust framework for intelligent energy management in various communication scenarios. This approach offers new insights into the systematic evaluation and intelligent distinction of energy-saving techniques in modern telecommunications.