It has been shown that dexmedetomidine (Dex) could attenuate postoperative cognitive dysfunction (POCD) via targeting circular RNAs (circRNAs). Circ-Shank3 has been found to be involved in the neuroprotective effects of Dex against POCD. However, the role of circ-Shank3 in POCD remains largely unknown. Reverse transcription quantitative PCR (RT-qPCR) was performed to detect circ-Shank3 and miR-140-3p levels in lipopolysaccharide (LPS)-treated microglia BV-2 cells in the absence or presence of Dex. The relationship among circ-Shank3, miR-140-3p and TLR4 was confirmed by dual-luciferase reporter assay. Additionally, Western blot and immunofluorescence (IF) assays were conducted to evaluate TLR4, p65 and Iba-1 or CD11b levels in cells. In this study, we found that Dex notably decreased circ-Shank3 and TLR4 levels and elevated miR-140-3p level in LPS-treated BV2 cells. Mechanistically, circ-Shank3 harbor miR-140-3p, functioning as a miRNA sponge, and then miR-140-3p targeted the 3'-UTR of TLR4. Additionally, Dex treatment significantly reduced TLR4 level and phosphorylation of p65, and decreased the expressions of microglia markers Iba-1 and CD11b in LPS-treated BV2 cells. As expected, silenced circ-Shank3 further reduced TLR4, p65 and Iba-1 and CD11b levels in LPS-treated BV2 cells in the presence of Dex, whereas these phenomena were reversed by miR-140-3p inhibitor. Collectively, our results found that Dex could attenuate the neuroinflammation and microglia activation in BV2 cells exposed to LPS via targeting circ-Shank3/miR-140-3p/TLR4 axis. Our results might shed a new light on the mechanism of Dex for the treatment of POCD.