Due to the absence of safety of the antipyretics to patients with cardiovascular dysfunction, new targets to treat inflammation have been pursued. mPGES-1 is a promising target because its inhibition would not cause the side-effects related to COX inhibition. To identify novel inhibitors of mPGES-1, we developed a ligand-based pharmacophore model that differentiates true inhibitors from decoys and enlightens the structure-activity relationships for known mPGES-1 inhibitors. The model (four hydrophobic centers, two hydrogen bond acceptor and two hydrogen bond donor points) was employed to select lead-like compounds from ZINC database for in vivo evaluation. Among the 18 compounds selected, five inhibited the fever induced by LPS. The most potent compound (5-(4-fluorophenyl)-3-({6-methylimidazo[1,2-a]pyridin-2-yl}methyl)-2,3dihydro-1,3,4-oxadiazol-2-one) is active peripherally (i.v.) or centrally (i.c.v.) (82.18% and 112% reduction, respectively) and reduces (69.13%) hypothalamic PGE production, without significant COX-1/2 inhibition. In conclusion, our in silico approach leads to the selection of a compound that presents the chemical features to inhibit mPGES-1 and reduces fever induced by LPS. Furthermore, the in vivo and in vitro results support the hypothesis that its mechanism of action does not depend on COX inhibition. Hence, it can be considered a promising lead compound for antipyretic development, once it would not have the side-effects of COX-1/2 inhibitors.