The aim of this research is to monitor and classify the broken chip signals from the dynamic cutting forces, in order to predict the surface roughness during the computer numerical control turning process utilizing the Meyer wavelet transform to decompose the dynamic cutting forces. The dynamic cutting forces of the broken chips and the surface roughness can be decomposed into the different levels. The levels of decomposed cutting forces can aid to explain the broken chip formation and the surface roughness profile in both time and frequency domains. The experimentally obtained results showed that the surface roughness frequency occurs at the higher level of decomposed cutting forces, especially at the fifth level, although the cutting conditions are changed. However, the chip breaking frequency appears at the lower level, which depends on the cutting conditions and the chip length. The ratio of the fifth level of decomposed feed forces to that of main forces is proposed to predict the surface roughness during the in-process cutting. It is understood that the broken chip formation can be separated clearly and the surface roughness can be predicted well during the cutting, regardless of the cutting conditions.