The representation for the Moore-Penrose inverse of the matrix ■ is derived by using the solvability theory of linear equations,where A∈Cm×n,B∈Cm×p,C ∈ Cq×n and D ∈ Cq×p,with which some special cases are discussed.