A general small-signal model for amplitude-, phase-, and polarization-to-intensity conversion in optical systems affected by chromatic dispersion, polarization-mode dispersion (PMD), and polarization-dependent loss (PDL) is presented, which extends a previous scalar model by Wang and Petermann . The model leads to simple intensity filters, which can be expressed as a linear combination of the components of the Stokes’ vector of the signal input state of polarization (ISOP), and facilitates the prediction of the ISOPs, which minimize/maximize the intensity modulation on the output signal. The model is first used to study the output intensity in a first-order PMD-compensated single-channel system with either input amplitude, or phase, or polarization modulation. The small-signal model provides a good prediction of the received intensity up to modulation indexes of about 20%–30%, according to the modulation type. The model is then successfully used in a semianalytical bit-error rate (BER) evaluation method to estimate the system penalty induced by cross-phase modulation (XPM) in a two-channel wavelength-division-multiplexed (WDM) dispersion- managed system with PMD compensation.