Memantine, a noncompetitive NMDA receptor antagonist with neuroprotective properties, has been used for the treatment of Alzheimer's disease (AD). Administration of memantine to various transgenic AD mice has been reported to improve cognitive deficits, very often completely back to normal wild-type control levels. However, such great benefits of memantine in preclinical studies do not translate into clinical results of this drug, showing only marginal and transient efficacy in moderate to severe AD. To further address in vivo efficacy, we compared the effects of memantine at different disease stages in 5XFAD mice, one of the rapid-onset and most aggressive amyloid models. Specifically, we administered memantine once daily for 30 days to 5XFAD mice, which showed moderate (6–7 months of age) and robust (12–15 months) β-amyloid (Aβ) accumulation. Treatments with memantine (10 mg/kg, i.p.) reversed memory impairments in the younger 5XFAD mice, as tested by the contextual fear conditioning and spontaneous alternation Y-maze paradigms. Memantine had no effects on soluble Aβ oligomer or total Aβ42 levels in 5XFAD mouse brains. In contrast, subchronic treatments with memantine showed no behavioral benefits in the older 5XFAD group, which exhibited more profound memory deficits concomitant with highly increased concentrations of Aβ as compared with those of the younger 5XFAD group. Since subchronic memantine at the higher dose (30 mg/kg) impaired memory performances in wild-type controls, we further tested acute administration of 50 mg/kg memantine, which was reported to enhance hippocampal adult neurogenesis and memory function. However, this treatment also failed to rescue memory deficits in 12–15-month-old 5XFAD mice. Collectively, our results demonstrate that cognitive benefits of memantine independent of Aβ reductions were no longer observed in the 5XFAD Alzheimer mouse model during advanced stages, which may be reflective of the limited efficacy of memantine in clinical settings.