Aluminum alloys are increasingly being considered the material of choice in electric vehicle (EV) manufacturing due to their lightweight characteristics. As multi-material manufacturing expands, weld quality monitoring becomes more important because of EVs’ high safety standards. By monitoring the weld quality, inspecting time can be saved and it tends to replace non-destructive testing as well. Proper interface width in overlap joint configuration ensures stable weld strength. Especially, the interface widths are known to directly influence the mechanical properties during interfacial failures. Therefore, predicting the interface width accurately is critical during the welding property monitoring process. In this study, we present a method that is able to predict interfacial width in overlap joint configuration for laser welding of aluminum alloy using sensors and a CNN-based deep learning model. A spectrometer and a CCD camera were attached to the head. The inputs for the multi-input CNN-based deep learning prediction models were spectral signals, represented by light intensity through the spectrometer, and the molten pool's dynamic images filmed by a CCD camera. The output was the interfacial width collected from the cross-sectional analysis. Through this, we show high prediction accuracy in predicting the interfacial width in overlap joint configuration for laser welding of aluminum alloy.