This paper describes the results of a comparative investigation for establishing a sensitive and reliable method for on-line end mill wear and breakage detection using two sensors and different signal processing schemes. Spindle vibration in the direction of feed velocity and the feed force were the variables monitored. The RMS values of the third difference of the spindle acceleration, and the Residual Sum of Squares (RSS) of an Autoregressive Moving Average (ARMA) model fitted to the feed force and spindle acceleration have been found to be very sensitive to end mill wear and breakage. Based on these findings, two strategies for on-line detection of end mill wear and breakage have been proposed, one based on the RMS values of the third difference of spindle acceleration and the other on the RSS values of the feed force. Block diagrams of the corresponding computation algorithm for on-line implementation of the strategies are presented.