Abstract: A line spectrum detection method using Doppler-warping transform is proposed to detect high-speed airborne sound sources from underwater. The method firstly uses Doppler-warping transform to linearize the Doppler phase of the target line spectrum, then extracts the linearized target line spectrum points, and combines beamforming algorithm to achieve target detection. Numerical simulation and the sea trial data processing show that compared with the method of directly extracting the target line spectrum to achieve detection, the proposed method has the following advantages: (1) it can linearize the nonlinear phase of the target line spectrum produced by the Doppler effect, making the line spectrum tracking more stable and continuous; (2) the target line spectrum energy of the broadened frequency band is refocused to the vicinity of the original frequency, and the target line spectrum can be better extracted in a low signal-to-noise ratio environment; (3) non-linearization of the instantaneous frequency of the interference line spectrum can better realize the separation of the interference line spectrum. The sea trial data processing results show that the algorithm has achieved effective underwater detection of moving targets in the air as far as 16.8 kilometers away, and the detection ability and anti-interference ability have been significantly enhanced.