The following paper presents an online model-based monitoring algorithm that identifies and quantifies the level of waviness caused by grinding chatter exploiting accelerations measured on the machine. A Weighted Least Squares method has been applied in the frequency domain that leads to a clear interpretation of the forcing effect caused by the waviness at particular frequency components exciting the machine structure. A sensitivity analysis has been carried out in order to define the limits of applicability and the precision of the approach in relation to the adopted analytical models. The development has been supported by simulations produced by an integrated model implemented in Matlab/Simulink® environment. Finally, experiments on the real machine confirmed the good estimation performance of the algorithm.