The central cholinergic nervous system plays an important role in cognition, with acetylcholine hypofunction considered to be a major factor of dementia. Botulinum toxin type A (BoNT/A), a potent poison secreted by Clostridium botulinum, is used widely for dystonia treatment and facial cosmesis. BoNT/A injection inhibits acetylcholine release in the neuromuscular junction through cleavage of synaptosomal-associated protein of 25 kDa in cholinergic terminals. Furthermore, beyond the injection site, BoNT/A undergoes retrograde transport and transcytosis to the central nervous system from peripheral cholinergic terminals. However, whether peripheral BoNT/A injection affects the function of the central nervous system and induces learning deficits remains unclear. We injected mice with different doses of BoNT/A (2, 10, and 50 U/kg) or sterile saline (control) into the left whisker pad to test spatial learning performance at different times after injection using the Morris water maze. At 3 days and 4 weeks after injection, the spatial learning ability of the control and BoNT/A-treated mice showed no significant differences. Surprisingly, however, rather than spatial learning impairment at 6 weeks after injection, BoNT/A-treated mice spent less time than control mice in locating the experimental platform, indicating that BoNT/A facial injection might promote spatial learning. Furthermore, our study suggests that facial application of BoNT/A is safe and could play a positive role in ameliorating the spatial learning deficits associated with neurodegenerative diseases.