A new online grinding wheel wear monitoring approach to detect a worn out wheel, based on acoustic emission (AE) signals processed by discrete wavelet transform and statistical feature extraction carried out using statistical features such as root mean square and standard deviation for each wavelet decomposition level and classified using tree based knowledge representation methodology decision tree C4.5 data mining techniques is proposed. The methodology was validate with AE signal data obtained in Aluminium oxide 99 A(38A) grinding wheel which is used in three quarters of majority grinding operations under different grinding conditions to validate the proposed classification system. The results of this scheme with respect to classification accuracy were discussed.