Abstract: In view of the complex composition of energy consumption of CNC machine tools, it is difficult to predict energy consumption with high accuracy in theoretical analysis. A data-driven BP-Adaboost CNC machine tool consumption prediction model is proposed. This model introduces the ability of the Adaboost algorithm to integrate strong predictors, and improves the BP neural network. By repeatedly adjusting the weight of the BP weak predictor and the sample weight, a strong predictor is obtained, thereby improving the prediction accuracy. Experimental results show that the BP-Adaboost prediction model can predict the energy consumption of material cutting more accurately than the independent BP neural network prediction model, and the root mean square error and absolute error are reduced. It can be seen that the prediction model has practical feasibility in predicting the energy consumption of CNC machine tools, and provides a new tool support for the prediction of the total energy consumption of machine tools.