Abstract: Objective To investigate the effect of intracerebroventricular injection of lipopolysaccharide (LPS) on cognitive function and establish inflammation-induced cognitive dysfunction model in mice. Methods Forty C57 mice were randomly divided into four groups:control group, 10 ng group, 100 ng group, 2 000 ng group. Intracerebroventricular injection of LPS at dose of 0, 10, 100, 2 000 ng was performed after 5 days of acquisition training in a Morris water maze (MWM). The probe test was conducted 1 day after LPS injection, and working memory was tested on day 1 to 3. The other 40 C57 mice were divided as the same way and used for novel object recognition test. Intracerebroventricular injection of LPS at the same doses was performed after acquaintance and training and mice were tested 1 day after injection. Ten more mice were injected with LPS at dose of 0, 2μg for IL-1β and TNF-α examination in hippocampus. Results The Morris water maze and novel object recognition test indicated that cognitive declined in all LPS groups, while 2 000 ng LPS group showed decrease of platform-site crossovers (72%), distance around platform (46%), percentage of distance travelled in the target quadrant (32%) during probe testing, increase of latency to the platform by twice during the working memory test in Morris water maze. Percentage of time and counts for novel object exploration in novel object recognition test were decreased by 22% and 21%, respectively. The ELISA test showed an increase of IL-1β (55%) and TNF-α (41%) levels in the hippocampus of mice in 2 000 ng LPS group. Conclusion Intracerebroventricular administration of lipopolysaccharide at dose of 2 μg induces cognitive decline in C57 mice, which suggests that LPS-induced CNS inflammation can contribute to establish a cognitive dysfunction model in mice.