In conventional mold and die manufacturing, the machining process is conducted in the annealed state (lower hardness of steel) followed by heat treatment, electric discharge machining (EDM), and finishing by surface grinding or polishing. Using high-speed machining (HSM), it is possible to perform the machining in the hardened state, eliminating the heat treatment and EDM at the middle stage, and even perform a better surface quality, thus reducing the effort for grinding or polishing. However, in HSM, tool wear is a serious issue and requires a robust monitoring system to avoid defective products. Furthermore, there is a challenge in tool monitoring for this application where the depth of cut is very small with a high cutting speed, thus difficult to capture the cutting process through a sensor compared to conventional machining. This work is addressed to conduct feasibility studies of using vibration sensors to monitor tool wear progress. Five positions of sensors are investigated: close to the workpiece, Y-axis slideways, spindle housing, and two direct motors for workpiece rotation. The signals are extracted into several features; maximum, root mean square (RMS), mean, standard deviation, kurtosis, skewness (in the time domain), and peak spectral amplitude (in the frequency domain). The best feature is selected using Pearson correlation analysis, then plotted together with the tool flank wear length data to confirm the capability of tool wear monitoring. The results showed that the tool wear progress can be monitored using the RMS value of signals from a sensor embedded close to the workpiece.