Background. Alzheimer’s disease (AD) is a common health condition affecting senile people and leads to severe cognitive dysfunctions. Acupuncture has been shown to be a possible alternative natural remedy for AD in some animal studies. Objective. To perform a systematic review to identify the effect of verum-acupuncture compared with sham-acupuncture on learning and memory performance among animal models of AD. Methods. Experimental animal studies of treating AD via verum- and sham- acupuncture were searched in nine electronic databases, including Sciverse ScienceDirect, PubMed, Springer, Ebsco Medline, AMED, EMBASE (Elsevier), Scopus (Elsevier), PsycINFO (ProQuest), and OVID from the dates of the databases’ inception to June 2019. The Morris water maze test was considered as an outcome measure. The software Revman 5.3 and Stata 16.0 were used to conduct the meta-analysis. Heterogeneity was examined by using I2 statistics. The publication bias was assessed via Begg’s test by Stata 16.0. Results. Twelve studies involving 229 animals met the inclusion criteria. Most of the studies had a moderate quality according to SYRCLE's risk of bias tool for animal studies. The results of the meta-analysis indicated that verum-acupuncture could reduce the escape latency (MD = −12.90, 95% CI (−17.08, −8.71), p<0.001) and increase the time spent in the original platform quadrant (MD = 7.28, 95% CI (4.23, 10.33), p<0.001) and frequency of the crossing former platform (MD = 2.01, 95% CI (1.53, 2.50), p<0.001) compared with the sham-acupuncture. Conclusions. Acupuncture is effective in improving cognitive functions in AD animal models, and this benefit is more than just a placebo effect. Further clinical trials are needed to confirm the findings.