This paper presented a new technology, which was the online wheel wear evaluation in grinding WC-Co ceramic coating by using wavelet packet decomposition. On the basis of the measurement methods, theory of online wear evaluation of wheel was introduced. And the selection principle of characteristic signals was brought forward, the grinding acceleration signal was the best choice among many characteristic signals. The time domain signal analysis method of wavelet packet decomposition was analyzed. Lastly, experiments on the wear of wheel were performed, and the experimental results showed the validity of the method.