Abstract: Time-frequency distribution (TFD) is the best analysis tool for the single-component FM signals, however, it has cross-term interference in the analysis of multi-component LFM signals. For suppressing cross-term interference and keeping some time-frequency local characteristics, parameters of TFD were optimized based on the time-frequency information entropy, and values of some parameters were recommended. Simulation tests showed that the method improved the resolution of TFD. Research results were applied to monitor bearing wear of the internal combustion engine, and fault features were extracted and the state change of internal combustion engine was revealed.