In view of the problem that the surface roughness of the hole can not be detected in machining in realtime, the acoustic emission (AE) nondestructive testing technology is applied to the processing technology of the fluid magnetic abrasive. Through the analysis and study of the collected AE signal, it was found that the AE signal has a significant correlation with the surface roughness of the workpiece. The characteristic value of the energy count AE was extracted by the eigenvalue parameter analysis method, and the distribution of the energy count could be used to characterize the change of the surface roughness of the workpiece. This study result provides a valuable reference for realtime dynamic detection of fluid magnetic abrasive.