In biomimetic polarized-light navigation, atmospheric polarization modes can be used as directional benchmarks to achieve precise navigation and location. In the measurement of atmospheric polarization modes, when the direct sunlight lens or field of view is obstructed, the measurement will be interfered, thus resulting in reduced navigation accuracy. To reduce the effect of these perturbed areas on polarization-navigation accuracy, this study proposes an anti-interference algorithm based on the polarization angle and degree to eliminate interference and uses the Rayleigh scattering model to obtain the direction of the solar meridian for navigation orientation. Experimental results show that this method can not only accurately partition the interference area but also improve the accuracy of polarization navigation methods. For data with heading-angle errors exceeding 2.00° before and after anti-interference, the average heading-angle errors are 5.37° and 1.56°, respectively. For data with heading-angle errors of less than 2.00° before and after anti-interference, the average heading-angle errors are 1.13° and 0.59°, respectively.