The ship-radiated noise possesses a stable part of the line-spectrum at low frequency range, containing rich information, and it is meaningful for positioning and distinguishing underwater targets. In order to separate the ship-radiated noise line-spectra from the broadband background noise effectively, the method for analyzing the ship-radiated noise signals by using Adaptive Line-spectrum Enhancement (ALE) algorithm is put forward. As a variant of adaptive Finite Impulse Response (FIR) Digital Filter (DF), ALE is a typical application of adaptive noise canceller. The algorithm is able to find and match the line-spectrum signals and separate them from wide band addictive noise without reference signals. Then, the line-spectrum signals could be enhanced. By theoretical analysis simulation and actual ocean-trial, the research on the Low Frequency Acquisition and Ranging (LOFAR) spectrum shows that ALE algorithm is able to meet the expectation competently. Comparing with conventional methods, the detection performance enhances observably, and radiated noise line-spectra are able to be effectively extracted.