Research on the identification and diagnosis of high-frequency friction signals under abnormal contact conditions of me? chanical seals is limited. Therefore， a seal monitoring test-bed is established using high-frequency， wide-range， and highly sensi? tive acoustic emission （AE） technology. Acoustic emission signals are collected under various operating conditions of the mechani? cal seal， including start-stop， no pressure， low pressure， high pressure， low speed， and high speed. Then the collected data are processed and analyzed using time-domain， frequency-domain and time-frequency-domain analysis methods. A method for deter? mining the interface friction state of the mechanical seal， based on the root mean square （RMS） change of the AE signal， is estab? lished. This method reveals the corresponding relationship between the end-face friction signal and the running state of the mechani? cal seal， from contact to non-contact， and further proves the seal operation process. The research identifies the evolution of the fric? tion signal during the transition from boundary lubrication （BL） and mixed lubrication （ML） in end-face contact friction to sliding contact signals under non-contact hydrodynamic lubrication （HL）. The frequency band of the AE signal during seal interface fric? tion is found to be between 240 kHz and 320 kHz. Through time-frequency analysis of the AE signal， it is shown that continuous wavelet transform （CWT） effectively represents the time-frequency characteristics of the AE signal under different working condi? tions of the mechanical seal. The research results of this paper provide a theoretical foundation and data support for research on me? chanical seal condition monitoring and fault diagnosis.