Given two orientations $Q, Q'$ of a Dyinkin diagram. Is it always true that after a sequence of mutations, $Q$ becomes $Q'$? Are the some references about this? Thank you very much.
Yes, it's true for any two acyclic quivers with the same underlying diagram having different orientations. See lectures notes on cluster algebras by Robert Marsh.