Infectious bursal disease virus (IBDV) is of major economic importance to the poultry industry. The global population of IBDV strains was recently classified into nine genogroups based on the amino acid sequence of the hypervariable region (HVR) of the capsid protein (VP2).  However, how the genogroups relate to each other antigenically is poorly understood. To address this, we inoculated groups of chickens with a panel of seven recombinant IBDVs that contained HVRs from six different genogroups from Europe, South America, China, Australia, and the USA. We obtained the serum, conducted a cross-neutralization assay, and applied antigenic cartography analysis to the results. Sequence distance (as measured by amino acid substitution count) correlated weakly with antigenic distance (Pearson’s Coefficient = 0.467, R2 =0.218), suggesting that the antigenic relationships of IBDV strains could not be accurately predicted based on the HVR sequence alone. We hypothesized that classifying IBDV based on the structural similarity of the capsid might lead to a better correlation with antigenic distance. To test this hypothesis, we used Alphafold3 to predict the structure of the VP2 timers of the seven recombinant strains and we developed a computational algorithm to quantitatively compare the structures and assign an overall “structural distance” score for each strain. The structural distance score produced by our algorithm showed a substantially improved correlation with the observed antigenic distance (Pearson’s coefficient = 0.784, R2 =0.614). In the future, it might be possible to classify IBDV strains into a “structural phylogeny“, to better inform vaccine antigen selection.