Recently, a series of highly effective peptide- or protein-based HIV fusion inhibitors have been identified. However, due to their short half-life, their clinical application is limited. Therefore, the development of long-acting HIV fusion inhibitors is urgently needed. Here, we designed and constructed a protein-based, long-acting HIV fusion inhibitor, termed FLT (FN3-L35-T1144), consisting of a monobody, FN3, which contains an albumin-binding domain (ABD), a 35-mer linker (L35), and a peptide-based HIV fusion inhibitor, T1144. We found that FLT bound, via its FN3 component, with human serum albumin (HSA) in a reversible manner, thus maintaining the high efficiency of T1144 against infection by both HIV-1 IIIB (X4) and Bal (R5) strains with IC50 of 11.6 nM and 15.3 nM, respectively, and remarkably prolonging the half-life of T1144 (~27 h in SD rats). This approach affords protein-based HIV fusion inhibitors with much longer half-life compared to enfuvirtide, a peptide-based HIV fusion inhibitor approved for use in clinics. Therefore, FLT is a promising candidate as a new protein-based anti-HIV drug with an improved pharmacokinetic profile.