Abstract: In order to solve the influence of tool wear and edge collapse on machining quality and efficiency during machining, a tool wear monitoring system based on EMD-PSO-HMM through robot learning technology is designed. Firstly, the Spindle current signal under different tool wear states is extracted. Due to the limitations of wavelet analysis and Fourier analysis in the signal analysis process, The EMD algorithm is used to decompose the spindle current signals at different scales during the machining process and extract feature parameters. The extracted feature values are entered into the HMM model for training and iteration. In order to solve the problem of local minimum values in the HMM model training process, Particle swarm optimization algorithm is introduced in this paper introduces to globally search for the input parameters of the HMM model, so it can achieve the optimal value. The tool wear monitoring system based on EMD-PSO-HMM have high accuracy in the actual tool wear status evaluation.