Energy consumption becomes a serious concern for manufacturing industry because it not only consumes substantial amounts of energy and materials but also produces huge amount of greenhouse CO2 emissions. With the drive for sustainable development, manufacturing companies are under increasing pressure from government regulations to reduce energy consumption and related emissions. Hard milling (i.e., milling of hardened steels) is an important precision cutting process. Compared with grinding, hard milling may produce superior surface integrity of precision components such as bearings, dies, and molds. However, the poor machinability of hard milling leads to very low material removal rates and thus low energy efficiency. Previous research has focused on the relationship between energy consumption and process conditions at the machine and spindle levels. However, little has been done to investigate the energy consumption at the process level (actual material removal). In addition, tool wear is inevitable in precision cutting. However, the effect of tool wear progression on energy consumption at machine, spindle, and process levels is yet to understand. In this study, power profile and energy consumption at the process level as well as the machine and spindle levels were characterized in hard milling of AISI H13 steel (50 ℗ł1 HRC). A new concept {esc}(3z{esc}(Bnet cutting specific energy{esc}(3y{esc}(B has been defined to investigate the energy consumed at the process level, i.e., by the actual material removal process. The relationships between cutting conditions and energy consumption at each level have been established. The results have shown that traditional empirical models may predict specific energy at the machine and spindle levels, but not net cutting specific energy at the process level. To solve the problem, a new power regress model has been developed to predict net cutting specific energy with high accuracy at the process level. The results have shown that tool wear has the greatest influence on net cutting specific energy compared with feed and cutting speed. A power regress model has been developed to predict net cutting specific energy with high accuracy. Material removal rate (MRR) is found not a unique identifier for net cutting specific energy. Energy efficiency increases with MRR. In addition, up milling consumes slight more energy than down milling. The emissions and environmental impact induced by energy consumed by the machine and embodied energy of the cutting tool have been investigated. The cutting tool embodied energy had a significant effect on the total specific energy, process emissions, and environmental impact in hard milling from the viewpoint of machining system. The predictive models have been developed to quantify the relationships between material removal rate, specific energy, emissions, and environmental impact.