The role of the humoral immune response in prevention against HIV-1 infection is still incompletely understood. However, neutralizing antibodies to certain epitopes on HIV-1 envelope glycoproteins inhibit HIV-1 infection in vitro and in vivo. Passive administration of these antibodies by themselves or in combination completely protected hu-PBL-SCID mice or macaques from intravenous, vaginal, as well as maternal-fetal mucosal transmission. All these studies provide direct experimental evidence that neutralizing antibodies are potent enough to prevent HIV infection, and strongly suggest that neutralizing-antibody-based vaccines could provide effective protection against HIV-1, despite the potent action of CTLs. Some neutralizing epitopes have been defined in vitro and in vivo. Unfortunately, none of the neutralizing-antibody-based candidate vaccines has been demonstrated to induce enough protective activity. Weak antigenicity and immunogenicity of neutralizing epitopes on native or recombinant proteins and other factors made it difficult to induce neutralizing-epitope-specific antibody responses in vivo enough to prevent against primary isolates. Recent studies indicated that HIV-1 variations resulted in escape from neutralization or the CTL responses, which may be the principal challenge for HIV-1 prevention. Epitope vaccine as a new strategy activating both arms of the immune system, namely, using the "principal neutralizing epitopes" and the CTL epitopes in combination, should provide new hope for developing an effective vaccine to halt the HIV-1 epidemic.