This paper proposes a seismic random noise suppression method using the predictive filter in the discrete cosine transform (DCT) domain, and then evaluates its performance. In comparison with the discrete Fourier transform (DFT), DCT can represent seismic signals with fewer coefficients, i.e., DCT has superior energy compaction for seismic signal. That is why DCT can separate seismic signals from random noise better. The results from both synthetic and real data show that the proposed method achieves better performance for random noise suppression and signal preservation compared with the f-x prediction filter.