In today’s world, energy efficiency is becoming increasingly crucial, due to its impact on sustainability in production. Designing systems that consume less energy and manage resources efficiently is essential. Variations in operating speed can affect processing time, energy consumption, idle times of subsequent machines, work delays, and missed deadlines. While most studies focus on prediction parameters like cut depth and cut area to estimate the energy consumption or processing time, our approach emphasizes variations in G-code motion parameters. To enhance both precision and the adaptability of the model to all Computer Numerical Control (CNC) machines in no-load condition, we propose an intelligent hybrid model that combines physical and data-driven approaches. The first approach employs curve fitting based on the physical model, while the second utilizes deep neural network (DNN) models optimized through hyperparameter tuning. The DNN topologies we evaluated exhibit high performance, with prediction errors of less than 1% for all models. Our approach provides a more precise and comprehensive understanding of energy consumption patterns in manufacturing processes, enabling manufacturers to make informed decisions about energy utilization, cost reduction, and sustainability. Our study aims to advance the field of energy efficiency in manufacturing and contribute to a more sustainable future.