The signal-to-noise conditions can affect the temporal correlation of low signal-to-noise ratio LiDAR echo signals, leading to a decrease in the accuracy of signal recognition results. In this regard, a self-adaptive recognition method for low signal-to-noise ratio LiDAR echo signals is proposed. Firstly, under low signal-to-noise ratio environment, an independent component analysis model is used to perform independent component analysis on the low signal-to-noise ratio echo signal, separating the source echo signal and interference signal, and obtaining the echo signal after interference suppression. Then, using short-time Fourier transform to obtain the time-frequency spectrum of the echo signal, the improved SSD network model is trained through the time-frequency spectrum to achieve adaptive recognition of low signal-to-noise ratio echo signals. Simulation analysis shows that the proposed method is feasible and effective in the field of adaptive recognition of low signal-to-noise ratio echo signals, which can effectively improve the reliability of LiDAR echo signals.