Abstract: Basic principles of target detection upon forward acoustic scattering are presented. An approach of direct blast suppression (DBS-AF) based on adaptive filtering is proposed to suppress the direct blast. The DBS-AF is extended to the linear frequency modulation (LFM) signal, where the envelope of the signal is regarded as a "general waveform" and then imported into the adaptive filter. Applying the DBS-AF method to the data collected from a lake trial yields an output detection curve, in which the direct blast is mapped to the background while the acoustic field aberration is represented by the peak-value fluctuation. The inhibitory effect on single channel is about -5 dB, which is then enhanced by exploiting the mean-value removal approach as a preprocessing technique. The direct blast is further suppressed to a level of -10 dB by making full use of multi-channel receptions. The main factors affecting the performance of the algorithm are: the fluctuation degree of the receptions during the weighting vector training period and the ratio of the forward scattered signal power to that of the direct blast when the target is present.