Tool condition monitoring (TCM) plays an important role in guaranteeing workpiece quality. There-
fore, it is meaningful to monitoring the tool wear condition in time. In this paper, a tool wear monitoring system based
on the heterogeneous ensemble learning model was proposed to overcome the limitation of the single classifier. In this
system, the SVM, RBF and HMM models were selected as base classifiers depend on the base classifier selection cri-
terion. In order to test the performance of the monitoring system, carbon fibre reinforced plastics (CFRP) drilling ex-
periment is carried out. Feature extraction technology in time domain is used for force and vibration signals, and the
LPP algorithm is used to realize the feature selection. By the comparison with ensemble learning and single classifiers,
it’s proved that the ensemble learning has better accuracy and stability.