An eigenstructure assignment-based method for reconfigurable control systems designs is developed. When the dynamics of the process being controlled are changed due to operating condition variations, or system component failures, the proposed method is capable of reconfiguring the control system so that the eigenvalues and eigenvectors of the original closed-loop system are recovered to a maximum extent. If full state feedback is permissible, the method guarantees the stability of the reconfigured system. If only system output is accessible, the method will recover the subsequent dominant eigenvalues and eigenvectors of the original system, and the stability of the reconfigured system will be maintained if a sufficient condition derived herein is satisfied. This condition unveils some intrinsic relationships between the degree of process variations, stability margins of the original system and the stability of the reconfigured system with output feedbacks. The steady-state performance of the original system under step inputs is also recovered via an optimal selection of a closed-loop system input weighting matrix. To illustrate the effectiveness of the proposed reconfigurable control system design technique, an aircraft longitudinal control system is used to demonstrate the reconfiguration procedure.