During seismic investigations, multiple and unexpected sources may cause serious interference on seismic records, and coherent noise generated by another unwanted active source could result in extremely poor data quality. Because airgun arrays have been widely used as the sound source in marine seismic surveys, the noise generated by another airgun array usually has similar characteristics to the primary signals in both frequency bands and wave forms, so the suppression of this type of coherent noise is very difficult. In practice, seismic crews try to avoid conducting multiple surveys simultaneously in a same area, so the source interference problem normally does not occur, and suppression of coherent noise from another active source has rarely been discussed and proposed before. This paper presents a dataset in which part of the records are contaminated by shot noise from another seismic vessel, and proposes a hybrid approach to suppress the coherent noise from that unwanted seismic source. Noise subtraction and primary signal preservation within different data properties are considered to begin the noise suppression. Based on different noise characteristics from various source directions and wave propagation paths, coherence noise can be separated from primary signals in frequency–wave number (F–K), frequency–time (F–T) and intercept time–slowness (tau–p) domains, respectively. This hybrid coherent noise suppression approach involves applying three different filters, F–K, F–T and tau–p, to the contaminated dataset. Our results show that most of the coherent noise generated by another seismic source could be suppressed, and seismic images could be substantially improved.