Antibodies are protein molecules indispensable for many therapeutic, diagnostic, and research purposes due to their exquisite ability to selectively recognize and bind a given antigen. The particular area of the antigen recognized by the antibody is called the epitope, and mapping of such epitopes can provide important mechanistic insights and indicate for what applications an antibody might be useful. Even though the nature of epitopes can be complex, epitopes in protein antigens are broadly grouped into linear or discontinuous epitopes depending on the positioning of the epitope residues in the antigen sequence and the requirement of structure. Specialized methods for mapping the two different classes of epitopes have been developed. While different in their detail, all of the experimental methods rely on assessing the binding of the antibody to the antigen or a set of antigen mimics. Classical epitope mapping methods, utilizing truncated proteins, small numbers of synthesized peptides, and structural analyses of antibody-antigen complexes, have been significantly refined. Current state-of-the-art methods involve combinations of mutational scanning, protein display, and high throughput screening in conjunction with bioinformatic analyses of large datasets. In line with the significant development of in silico protein structure prediction in recent years, computerized approaches for epitope prediction have also matured considerably.