Glycoprotein D (gD-2) is the entry receptor of herpes simplex virus 2 (HSV-2), and is the immunogen in the pharmaceutical industry's lead HSV-2 vaccine candidate. Efforts to prevent genital herpes using gD-2 subunit vaccines have been ongoing for 20 years at a cost in excess of $100 million. To date, gD-2 vaccines have yielded equivocal protection in clinical trials. Therefore, using a small animal model, we sought to determine if a live-attenuated HSV-2 ICP0− virus would elicit better protection against genital herpes than a gD-2 subunit vaccine. Mice immunized with gD-2 and a potent adjuvant (alum+monophosphoryl lipid A) produced high titers of gD-2 antibody. While gD-2-immunized mice possessed significant resistance to HSV-2, only 3 of 45 gD-2-immunized mice survived an overwhelming challenge of the vagina or eyes with wild-type HSV-2 (MS strain). In contrast, 114 of 115 mice immunized with a live HSV-2 ICP0− virus, 0ΔNLS, survived the same HSV-2 MS challenges. Likewise, 0ΔNLS-immunized mice shed an average 125-fold less HSV-2 MS challenge virus per vagina relative to gD-2-immunized mice. In vivo imaging demonstrated that a luciferase-expressing HSV-2 challenge virus failed to establish a detectable infection in 0ΔNLS-immunized mice, whereas the same virus readily infected naïve and gD-2-immunized mice. Collectively, these results suggest that a HSV-2 vaccine might be more likely to prevent genital herpes if it contained a live-attenuated HSV-2 virus rather than a single HSV-2 protein.