<p>The type I collagen derived from Nile tilapia (<i>Oreochromis niloticus</i>) contains amino acid sequences that are conducive to the generation of angiotensin I-converting enzyme inhibitory peptides (ACEIPs). This study employed an approach combining machine learning, in silico simulations and in vitro experiments to identify four novel ACEIPs. Small-sample quantitative structure-activity relationship and DLKcat models identified the four novel peptides as potential ACE inhibitors. Molecular docking and all-atom molecular dynamics simulations indicate that four peptides exhibit stable binding to ACE, with peptide PGPPGR exhibiting the strongest interaction. MM/PBSA analysis supported these findings. Weak interaction forces analysis indicated that the proline residue at the N-terminus facilitates hydrogen bond formation with active amino acids in ACE, a phenomenon attributed to its secondary amine configuration. At the same time, Lineweaver Burk plots demonstrated that all four peptides function as competitive inhibitors. In vitro assays validated the ACE inhibitory activity of the peptides, and the IC₅₀ values ​​of the four peptides were all in the millimolar range, among which IC₅₀ values of 0.88 ± 0.03 mM, 2.40 ± 0.08 mM, 4.64 ± 0.15 mM, and 4.83 ± 0.14 mM, for peptide PGPPGR, DGPR, PGPR and PGPK, respectively. Meanwhile, the ACE inhibitory activity exhibited by the peptide PGPPGR was comparable to that of the positive control peptide (IC₅₀ value of 0.58 ± 0.06 mM). This research presents an efficient screening methodology for the discovery of bioactive peptides and underscores the potential of fish by-products as ACE inhibitors.</p>