An analysis of chirp sonar signals backscattered from ocean sediments is presented in this paper to estimate sound attenuation in ocean sediments. The theory of applying a matched filter to linear frequency modulation chirp signals for generating time compressed pulses (CPs) is described and implemented. After the matched filter technique was applied, the signal-to-noise ratio (SNR) of the CP was enhanced. However, a smearing effect resulting from the convolution operation diminished the advantage of the improved SNR. The chirp signal processing algorithm is described and applied to a sample data set collected from an area with a variety of different sediments offshore northeastern Taiwan. For the exploration of signal features, time-frequency analysis is employed to reveal the characteristics of chirp signals presented in the joint time-frequency domain. Based on the spectral ratio method, ocean sediment attenuation properties are compared to those obtained from core samples analyzed in the laboratory. The observed attenuation coefficient was lower than those of some of the previous studies, but it was still within the reported data range. Factors that could contribute to the uncertainty are discussed.