The texturing of the tool surface can significantly improve the cutting performance of the tool. However, laser processing is characterized by rapid heating and quenching, which can lead to problems such as remelting layer stacking and microcracking. In order to solve the above problems, heat-assisted laser processing technology is introduced in this paper. Since titanium alloy is a difficult-to-machine material, the tool is subjected to large milling forces during the milling process, which leads to the dynamic response and vibration of the mechanical system, which in turn affects the tool life and the machined surface quality. Therefore, accurate prediction of the milling force can adjust the cutting parameters in time, ensure the machining quality at the same time, and make the milling force in a reasonable range, to improve the processing efficiency and reduce tool wear.In summary, this study takes the cemented carbide ball nose milling cutter as the research object, combines the heat-assisted process and laser processing technology, builds a milling test platform, and proposes a method based on the dung beetle algorithm (DBO) to optimize the variational mode decomposition (VMD) parameters, and combines the wavelet packet threshold noise reduction (WPT) method to denoise the original signal. The time-frequency analysis was carried out by using the Hilbert-Huang transform (HHT) to explore the variation of tool milling performance under different thermal auxiliary temperatures. On this basis, combined with Bayesian optimization (BO), convolutional neural network (CNN), bidirectional long short-term memory network (BiLSTM) and multihead-attention mechanism, a regression analysis model is formulated for real-time monitoring and prediction of milling force. Through verification, the R2 value of the model reaches 0.996 7 on the training set and 0.991 94 on the test set, which proves the accuracy of the model.This study proposes a new method for defect repair in the process of microtexture preparation and provides an effective method for the prediction of milling force in titanium alloy milling.