In the research study, there was applied a statistical model developed with the use of Principal Component Analysis (PCA) for the purpose of determining the impact of dipeptides’ structure on their antioxidant bioactivity. Sequences of 47 peptides were taken from a BIOPEP-UWM database of proteins and bioactive peptides (https://biochemia.uwm.edu.pl/biopep-uwm/). The  selected descriptors (also called attributes) to describe the physicochemical properties of amino acids present in a dipeptide chain were numerically expressed and implemented from open- access computer software or databases such as: Peptide Property Calculator, Biological Magnetic Resonance Data Bank, ProtScale, Molar Polarizability Values and ImMunoGeneTics. PCA was carried out using a Statistica software. The number of principal components that were significant for the interpretation of the impact of dipeptides’ structure on their antioxidant activity was  determined based on the percentage of cumulative variance. It was 79.9 % and it was in line with the four components. The first and the fourth component determined the impact of N-terminal amino acid residue on the antioxidant activity of dipeptide, whereas the second and the third one referred to the effect of C-terminal amino acid residue. By means of PCA it was shown  that a model dipeptide possessing antioxidant activity should be characterised by the presence of N-terminal amino acid with a cyclic or aromatic ring or amino acid residue with a non-polar  side chain. The C-end of peptide should be composed of proline, histidine, leucine or valine. The results obtained are consistent with the reference literature data that refer to the research on  relationship between structure and activity of antioxidative peptides, assessed using a multivariate regression analysis. Based on the research study performed, it was found that the  chemometric method applied might be useful when designing food-derived peptides, including those possessing the activity studied.