Graphical abstract COVID-19 is a pandemic disease caused by novel corona virus, SARS-CoV-2, initially originated from China. In response to this serious life-threatening disease, designing and developing more accurate and sensitive tests are crucial. The aim of this study is designing a multi-epitope of spike and nucleocapsid antigens of COVID-19 virus by bioinformatics methods. The sequences of nucleotides obtained from the NCBI Nucleotide Database. Transmembrane structures of proteins were predicted by TMHMM Server and the prediction of signal peptide of proteins was performed by Signal P Server. B-cell epitopes’ prediction was performed by the online prediction server of IEDB server. Beta turn structure of linear epitopes was also performed using the IEDB server. Conformational epitope prediction was performed using the CBTOPE and eventually, eight antigenic epitopes with high physicochemical properties were selected, and then, all eight epitopes were blasted using the NCBI website. The analyses revealed that α-helices, extended strands, β-turns, and random coils were 28.59%, 23.25%, 3.38%, and 44.78% for S protein, 21.24%, 16.71%, 6.92%, and 55.13% for N Protein, respectively. The S and N protein three-dimensional structure was predicted using the prediction I-TASSER server. In the current study, bioinformatics tools were used to design a multi-epitope peptide based on the type of antigen and its physiochemical properties and SVM method (Machine Learning) to design multi-epitopes that have a high avidity against SARS-CoV-2 antibodies to detect infections by COVID-19.