Unpredictable tool life and premature tool failure are major problems in micro-machining. In this study, the failure mechanisms of micro-end-mills were studied during the machining of aluminum, graphite electrodes and mild steel workpieces. Hundreds of machining operations were performed, and the pictures of cutting edges were taken with a scanning electron microscope to identify fatigue and extensive stress-related failure mechanisms. Also, the cutting force variation was monitored, i.e. the relationship between the utilization-related changes at the tool structure (wear), and the outcomes (increasing cutting force which means raising stress on the tiny shaft). Inspection of the cutting force variation patterns of large numbers of micro-end-mills indicated that tool failure occurs with chip clogging, fatigue and wear-related excessive stress depending on the characteristics of the workpiece. Two tool breakage prediction methods were developed by considering the variation of the static part of the feed direction cutting force. These methods used segmental averages and wavelet transformation coefficients. The accuracy of the proposed approaches were tested with experimental data and the agreement between the predictions and actual observations are reported.