The invention relates to a Lamb wave-based autonomous damage identification imaging method in the technical field of detection of mechanical structures, which comprises the following steps of: constructing a sensing network, and correcting time reversal-based damage index DI of each sensing path in the sensing network; then setting 45% of the maximum value in all the corrected damage indexes of all the sensing paths as a threshold, and using the threshold to judge the degree of influences on the sensing paths caused by damages; and finally carrying out weighted distribution treatment on the damage indexes of all the sensing paths, and further obtaining the probability value that the damages emerge at a coordinate point (x, y). By using the method, the flight time for extracting wave signals scattered by defects can be avoided, and the shortcoming of depending on reference signals can be overcome. Not only can the requirements on real-time property and on-line property be satisfied, but also the multiple damages can be accurately identified and positioned, and the autonomous non-destructive testing of a plate-shell structure can be further realized, thus the method has very important practical value in aerospace, buildings and other fields.