A simple method to measure and compensate polarization-sensitive optical components (retarders) is proposed, based on Stokes vector polarimetry. First, the Mueller matrix of an unknown polarization-sensitive element with arbitrary retardance and fast-axis orientation is obtained by means of a conventional measurement and calculation polarimetric procedure. Then a novel numerical technique is proposed, allowing to express the measured retarder’s Mueller matrix in terms of retardance δ and fast axis orientation θ. Further, a method for calculating polarization compensators is proposed, to balance phase differences introduced by one or more retarders. Compensation is achieved by a single compensating retarder whose parameters are derived from the original retarder’s Mueller matrix, or its equivalent (δ,θ) pair. The technique works for all possible fast-axis orientations of the original retarder. The method was validated using an electrically controllable liquid crystal retarder as a compensator. The approach described here is constrained to optical elements that behave as nearly pure retarders, i.e. their diattenuation and depolarization are negligible. Possible applications are in the design of polarization-sensitive systems (biomedical, environmental, chemical, optical communications, etc.) where recovery and/or precise control of the polarization state of light is needed, in polarization tracing models based on the Mueller matrix formalism, and others.