Monitoring a micromachining process is essential to ensure machined surface quality. Vibration signal has been used to monitor conventional process, however, to apply this signal to monitor a micromachining process is not a simple task. One of the main difficulties is to identify which signal features is correlated to changes in the micromachining process, such as roughness. This work investigates the correlation between vibration signals for an end micromilling operation. Different features were extracted from time and frequency domain aiming to analyze their correlation to the roughness and identify those better correlated to surface roughness. Six carbide end mill tools with a diameter of 0.4 mm were tested under different parameters. Vibration signals were collected, and roughness was measured for each machined channel. Through the analysis of the vibration signals and Pearson Correlation, features that are directly correlated with the roughness were identified, with emphasis on 2 x IPF (Insert Passing Frequency). The results show that the proposed features can be used for indirect monitoring of the end micromilling process, avoiding unnecessary stops and fabrication of rejected parts.