We introduce the notions of metric mean dimension of flows on the whole space and on subsets. Firstly, we develop Lindenstrauss-Weiss's classical inequality [LW00] for mean dimension of flows, and then we establish variational principles for metric mean dimension of general flows on subsets in terms of lower Brinkatok local entropy. We finally establish variational principles for metric mean dimension of uniformly Lipschitz flows on the whole phase space in terms of Rényi information dimension, Shapira's entropy, upper Brin-katok local entropy, Katok's entropy as well as general flows in terms of local entropy function.