ABSTRACT High-speed milling (HSM) aims to create high-quality surfaces and complex profiles and to increase the ability to eliminate excess material. This affects total energy consumption (EC), which in turn has an impact not only on the environment but also on production costs. Therefore, accurately evaluating how much energy is consumed by HSM is very important. EC is affected by stochastic tool wear, that is, when the latter increases, it leads to a corresponding increase in cutting force and vibration, thereby increasing EC. Therefore, the nonlinear processes caused by tool wear are determined through singularity vibration analysis combined with consideration of the geometry of the cutting edge with the aim of providing a predictive model for real-time prediction of the cutting force. Based on the cutting force components, an EC calculation model has been established. In particular, a new hybrid algorithm based on back-propagation neural network and multi-objective particle swarm optimisation is developed to determine the optimal cutting parameters, which can help minimise the total EC. HSM experiments were conducted to confirm the accuracy and availability of the proposed online monitoring and optimisation model. The proposed optimisation method based on the improvement model can reduce EC by 10.49% compared to the empirical option method. The feasibility and effectiveness of the proposed model is also verified through experimental processes.