In the field of cylinder liner-piston ring (CL-PR) wear calculations, the traditional calculation models are difficult to use as a guide for wear detection. Therefore, a new analysis model of CL-PR wear based on energy dissipation theory is established in this paper. By using a mixed lubrication equation and simulation modification method, a new characterization formula of CL-PR wear is theoretically established. The maximum error between its calculated results and the simulation results is only 1.167%. The qualitative relationship between the wear volume and the frequency band energy of the acoustic emission (AE) signal is established based on the energy conversion relationship. The AE signal acquisition platform of CL-PR wear is built on the EV80 combustion engine, and the Bayes Shrink method is used to set the threshold of large-amplitude disturbance AE events for noise reduction. The wear signal is processed by wavelet packet energy spectrum decomposition, and the resulting wear energy spectrum can better characterize the change of wear characteristics. The BP (Back Propagation) neural network is trained with the characteristic parameters of the energy spectrum, and the classification accuracy of training and finally experimental results reach 98.6% and 100%, respectively, proving that AE technology can be applied to CL-PR wear monitoring combined with the neural network.