A new method for simultaneous decoupling and assignment of eigenvalues of a linear square multivariable system is presented. Decoupling and pole-placement conditions are found in a structured way using the Faddeev-algorithm for computing adjoint matrices. The method allows pole-placement of all system poles in contrast to only those poles exceeding the number of transmission zeros in the case of classical decoupling methods. Further, the method avoids the internal instability associated with the application of the classical decoupling methods to non-minimum phase systems.