Methane is a powerful greenhouse gas. In Europe, methane emissions come mainly from breeding, and in particular from ruminants that have in their paunch a microbial ecosystem that ferments the plant matter. The purpose of this paper is to find the phenolic compounds of plants that could have an action on these microbes and limit the production of methane, in order to propose natural alternatives of feedstuff. As these compounds have a very wide variety of chemical structures, it is not possible to test them all. So we used data mining, and more specifically class association rules, to bring out compounds that could have a significant effect.