Aiming at generation of kerfs with groove and uplift profiles during laser scribing (LS) with nanosecond on Al5083, the relationships between two kinds of profiles and corresponding acoustic signals were investigated in this paper. A laser scribing experiment on Al5083 plate with nanosecond and ultraviolet laser was carried out, and microstructures of profiles was observed to explore the formation mechanism of kerfs with two different kinds of profiles, and acoustic emission (AE) signals were collected to analyze the differences. Wavelet Packet transformation (WPT) and support vector machine (SVM) analysis were performed to classify. The observation results illustrated the forming mechanism of uplift profile including the blocking of splash of molten metal and the generation of pores during solidification. AE signals analysis showed the variance and envelope area of WPT coefficients corresponding uplifts were significantly higher than those of grooves. Using WPT spectrum as feature vector after adding tags, SVM with Gaussian kernel was employed to classify. The accuracy reaching 92.57%, which verified the feasibility of WPT and SVM in the monitor of kerf profile during LS based on AE. This experiment provides a technical path for constructing LS monitor system based on AE.