We previously demonstrated that a kampo medicine, kakkonto, decreases lipopolysaccharide- (LPS-) induced prostaglandin E2 (PGE2) production by human gingival fibroblasts. In this study, we examined the herbs constituting kakkonto that exhibit this effect. Shokyo strongly and concentration dependently and kanzo and keihi moderately decreased LPS-induced PGE2 production. Shokyo did not alter cyclooxygenase-2 (COX-2) activity, cytosolic phospholipase A2 (cPLA2), annexin 1 and COX-2 expression, and LPS-induced extracellular signal-regulated kinase (ERK) phosphorylation. Kanzo inhibited COX-2 activity but increased annexin 1 and COX-2 expression and did not alter LPS-induced ERK phosphorylation. Keihi inhibited COX-2 activity and LPS-induced ERK phosphorylation but slightly increased COX-2 expression and did not alter cPLA2 and annexin 1 expression. These results suggest that the mechanism of shokyo is through the inhibition of cPLA2 activity, and that of kanzo and keihi is through the inhibition of COX-2 activity and indirect inhibition of cPLA2 activity. Therefore, it is possible that shokyo and kakkonto are clinically useful for the improvement of inflammatory responses.