I'm not a native speaker of French, I believe it should be "Mais ferme les maintenant." It is true that in spoken French, the word "ferme" would be just one syllable - the final e is silent. But in singing French songs, the final silent e is often pronounced in order to add another syllable when needed. So it is sung in two syllables as "ferm-e" where the final e is a sound not really found in English, kind of an English "schwa" sound but with fronted, rounded lips. Singing some kind of a soft "uh" sound is a good enough approximation. Both "fermer" and "fermez" would be ungrammatical and would be pronounced differently, like "ferm-ay."