Abstract: OBJECTIVE To explore the spatial recognition relationship between epitopes and antibodies based on the structure data of SARS-CoV-2 Spike protein antigen-antibody complex. METHODS Seven hundred and eighteen available SARS-CoV-2 antigen-antibody structural complexes were analyzed in multiple ways. Firstly, the epitope hotspots and the usage preference of antibody gene fragments were analyzed. Secondly, the spatial epitopes and CDR structures were extracted from the antigen-antibody binding interfaces, and clustering trees of the spatial epitopes and the antibody CDR structures were respectively constructed. The recognition relationship between the epitopes and the CDR structures was evaluated by comparing the similarity between the corresponding clustering trees. RESULTS The 94.02% epitope sites were mapped on RBD, and 4.44% on NTD, with the top10 hotspots being all located in RBM. The most common antibody VJ genes were identified as IGHV3-30/IGHJ4 and IGHV1-58/IGHJ3. The similarity between the clustering trees of spatial epitopes and those of the corresponding antibody CDR structures was significantly higher than that expected from random clustering trees, suggesting an intrinsic epitope-CDR matching. Further analysis of the data revealed that the epitopes with similar immunogenicity would be recognized by similar CDR structures, with CDR3 domain making the greatest contribution. CONCLUSION The current data identifies that the epitopes are concentrated in its RBD region while the IGHV3-30/IGHJ4 and IGHV1-58/IGHJ3 are the preferred VJ gene combination utilized for the production of Spike protein-targeting antibodies. It also suggests that similar epitopes are likely to be recognized by similar CDR structures. Collectively, these findings add a new theoretical basis for the SARS-CoV-2 antibody design and optimization.