I wonder why we use velocity inlet boundaries to incompressible flows, but mass flow at inlet for compressible flows? Could anybody guide me???

Typically the inlet mass flow for a problem is specified as it is usually known from physical data. For incompressible flow, since the density is constant and can be taken as unity, the mass flow rate reduces to the velocity. It should also be noted that most compressible solvers use a variable set where the mass flux is one of the dependent variables. Thus specifying the mass flux at the boundary is consistent with a Dirichlet boundary condition.