Abstract: Laser powder bed fusion (LPBF) technology, as a cutting-edge process in the field of metal additive manufacturing, has been successfully applied to high-end manufacturing fields such as aerospace. However, the strong coupling effect of multi-physical fields can easily cause dynamic instability of the molten pool, resulting in frequent porosity defects inside the workpiece, which seriously affects the stability of the forming quality. Traditional monitoring methods are limited by bottlenecks such as high cost and deployment difficulties, and it is difficult to meet the needs of industrial production. To this end, this paper proposes an online monitoring and quality intelligent discrimination method based on acoustic emission-deep learning fusion.An online monitoring system for the LPBF process based on acoustic emission sensing was developed. The mapping law between the acoustic emission signal characteristics and the forming quality was revealed by monitoring acoustic emission signals throughout the process, and the molten pool acoustic emission data containing more than 80,000 sets of samples was constructed. In order to solve the problem of extracting weak fluctuation features of the molten pool, we constructed a frequency domain feature extraction network based on the adaptive Fourier neural operator (AFNO) and a high-dimensional feature mapping classifier based on the Kolmogorov-Arnold network (KAN). We analyzed the dynamic characteristics of the molten pool through a multi-scale time domain feature fusion mechanism, and accurately mapped high-dimensional features with the help of high-dimensional manifolds, thus achieving enhanced characterization of weak fluctuation features in acoustic emission signals and high-precision quality discrimination. The experimental results show that the developed monitoring system can effectively capture the dynamic behavior of the molten pool, and the constructed discrimination model has a detection accuracy of more than 97%.