In this paper, we study mean‐field backward stochastic differential equations driven by G‐Brownian motion (G‐BSDEs). We first obtain the existence and uniqueness theorem of these equations. In fact, we can obtain local solutions by constructing Picard contraction mapping for Y term on small interval, and the global solution can be obtained through backward iteration of local solutions. Then, a comparison theorem for this type of mean‐field G‐BSDE is derived. Furthermore, we establish the connection of this mean‐field G‐BSDE and a nonlocal partial differential equation. Finally, we give an application of mean‐field G‐BSDE in stochastic differential utility model.