This study presents the development and application of a low-voltage portable atmospheric pressure discharge plasma (LV-PAPD) system coupled with optical emission spectroscopy (OES) for rapid classification of copper alloys. The proposed LV-PAPD source operates at atmospheric pressure and low driving voltage, enabling compact and low-voltage operation without the need for vacuum systems or complex gas handling. Characteristic neutral and ionic emission lines from a wide range of elements were clearly observed, confirming sufficient excitation capability under low-voltage conditions. When applied to Cu-based alloys, dominant Cu emission signals together with dopant-related emissions were identified. Using the regression method, the limits of quantification (LOQs) for Cu and Zn in these alloys were estimated to be approximately 59 wt.% and 23 wt.%, respectively. In addition, machine learning (ML) approaches substantially improve classification performance. Support vector machine (SVM) models achieved classification accuracies of up to 98%, while feature importance analysis from the random forest (RF) model reveals that both strong host-element emissions and weak dopant-related signals contribute significantly to alloy discrimination. The successful differentiation of copper alloys confirms the LV-PAPD-OES system's capability as a portable screening tool for rapid classification of copper alloys.