Despite significant progress in the development of diagnostic methods for influenza, avian influenza (AI) infection continues to represent a substantial threat to human health. Among the subtypes of AI, H5 influenza is highly infectious to animals and humans; however, there are no reliable H5 subtype-specific diagnostic systems owing to a scarcity of H5 subtype-specific detection elements. In this study, a new peptide aptamer (P1:KASGYTFTSF) was developed to recognize the H5 viral subtype using an in silico bioinformatics approach for predicting complementarity-determining regions (CDRs), and the aptamer was evaluated by immunoassays. The three-dimensional structure of influenza hemagglutinin (HA) and the peptide were used in a molecular docking study, and the peptide was compared to the epitope-derived peptide aptamer (H5-P2:KPNGAINF). Interactions between the peptides and the virus were then assessed by fluorescence-linked sandwich immunosorbent assay (FLISA), immunofluorescence staining assay (IFA), and rapid fluorescent immunochromatographic assay (FICT). P1 and H5-P2 both significantly detected H5N3 at 15.6 HAU/mL (P < 0.05), and P1 detected the virus more effectively (P < 0.05), consistent with the docking result. An optical image of the peptide recognizing an H5N3-infected cell was acquired by IFA, and was consistent with the antibody-linked IFA result. FICT employing the peptide showed the ability for H5 subtype-specific diagnosis, with 2-fold higher performance than that of a conventional, antibody-linked rapid test. This work shows the potential of a CDR-predicted peptide aptamer as a probe for immunological assays that can specifically recognize AI virus.