Energy consumption modeling of CNC machine tools is a prerequisite for manufacturers to save energy, reduce environmental pollution, and increase profits. It is of great significance to the healthy development of manufacturing. This paper aims to provide a power model suitable for predicting energy consumption of processing multiple materials in end milling. Firstly, this paper summarizes some limitations in the existing empirical power model of machine tool in cutting state (e.g., the applicability of predicting cutting energy consumption for various materials, the ease of obtaining power model coefficients, and the prediction accuracy of a model). Secondly, the cutting energy is analyzed from the perspective of mechanical mechanics. Combining the empirical and theoretical models analysis, key variables that affect the energy consumption in cutting status were determined. Then, an improved cutting state power model is proposed. The proposed power model is suitable for processing multiple materials in end milling. Thirdly, through experimental research, the effect of key variables on cutting power is discussed, especially its influence on the elastic-plastic term, velocity term, and thermal softening term in the calculation of equivalent shear stress. Suggestions for selecting different empirical power models are also discussed. Finally, the proposed model was verified, and the energy consumption of the end milling process was predicted. The proposed model improved applicability, accuracy, and rapidity to a certain extent of existing empirical models.