As metabolism impacts the efficacy and safety of therapeutic peptides and proteins (TPPs), understanding of the metabolic fate of TPPs is critical for their preclinical and clinical development. Despite the continued increase of new TPPs entering clinical trials, the metabolite identification (MetID) of these emerging modalities remains challenging. In the present study, we report an analytical workflow for MetID of TPPs. Using insulin detemir as an example, we demonstrated that top-down differential mass spectrometry (dMS) was able to distinguish and discover metabolites from complex biological matrices. For structural interpretation, we developed an algorithm to generate a complete and nonredundant theoretical metabolite database for a TPP of any topology (e.g., branched, multicyclic, etc.). Candidate structures of a metabolite were obtained by matching the monoisotopic mass of a dMS feature to the theoretical metabolite database. Finally, the MS/MS sequence tags enabled unambiguous characterization of metabolite structures when isobaric/isomeric candidates were present. This platform is widely applicable to TPPs with complex structures and will ultimately guide the structural optimization of TPPs in pharmaceutical development.