Spinocerebellar ataxia type 1 (SCA1) is a debilitating neurodegenerative disorder of the cerebellum and brainstem. Memantine has been proposed as a potential treatment for SCA1. It blocks N-methyl-D-aspartate (NMDA) receptors on neurons, reduces excitotoxicity and decreases neurodegeneration in Alzheimer models. However, in cerebellar neurodegenerative diseases, the potential value of memantine is still unclear. We investigated the effects of memantine on motor performance and synaptic transmission in the cerebellum in a mouse model where mutant ataxin 1 is specifically targeted to glia. Lentiviral vectors (LVV) were used to express mutant ataxin 1 selectively in Bergmann glia (BG). In mice transduced with the mutant ataxin 1, chronic treatment with memantine improved motor activity during initial tests, presumably due to preserved BG and Purkinje cell (PC) morphology and numbers. However, mice were unable to improve their rota rod scores during next days of training. Memantine also compromised improvement in the rota rod scores in control mice upon repetitive training. These effects may be due to the effects of memantine on plasticity (LTD suppression) and NMDA receptor modulation. Some effects of chronically administered memantine persisted even after its wash-out from brain slices. Chronic memantine reduced morphological signs of neurodegeneration in the cerebellum of SCA1 model mice. This resulted in an apparent initial reduction of ataxic phenotype, but memantine also affected cerebellar plasticity and ultimately compromised motor learning. We speculate that that clinical application of memantine in SCA1 might be hampered by its ability to suppress NMDA-dependent plasticity in cerebellar cortex.