The spectrum and characteristic values of acoustic emission signals of single diamond grit scratching alumina and zirconia ceramics at different cutting depths were studied experimentally. A time series Auto Regressive Moving Average (ARMA) model was established for the acoustic emission of diamond grit scratching alumina and zirconia ceramics by appropriate time series model and model order. The main frequency of single diamond grit scratching on alumina and zirconia ceramics is dominant on the frequency band of 0~110 kHz．The results show that the main frequency band energy of scratch alumina Acoustic Emission (AE) signal increases with the increase of cutting depth，and the maximum amplitude of AE signal is at a higher frequency in the main frequency band. The main frequency band energy of scratch zirconia Acoustic Emission (AE) signal decreases with the increase of cutting depth，and the maximum amplitude of AE signal is at a lower frequency in the main frequency band. The eigenvalues of the acoustic emission signal will increase with the increase of the depth of cut. The appropriate ARMA model has a high similarity with the acoustic emission signal of diamond grit scratching alumina and zirconia ceramics，and the linear relationship between the characteristic value and the cutting depth is also good. It shows that the appropriate ARMA model can better characterize the acoustic emission signal characteristics of single diamond grit scratching alumina and zirconia ceramics. The real time analysis of scratching depth of single diamond grit can be done by use of the appropriate ARMA model.