The correction of flat fields and polarization offset is a necessary step in the reduction of the full-disk solar observing data. The general used methods of the measurement of flat fields are summarized as follows. The first one is realized by randomly moving the solar telescope and making an average of thousands of observed images of the quiet region near the disk center, which is adopted by the large aperture solar telescopes with a small field-of-view. The second one is to shift the solar images with slightly different pointings, which is mostly used in the full-disk solar images. The third method is to use a diffuser to generate the uniform light source and it can be used in the full-disk solar images, too. Observing the nearby continuum of the selected working spectral line is generally used to obtain the polarization offset. It is worth to mention that, all these methods need special calibration observing mode and occupy the valuable scientific observation time. In the paper, we propose a new method to derive the flat fields and polarization offset from the routine observing data. Assuming that the flat fields and polarization offset do not change during a solar rotation, we firstly got the formula which can be used in the calculation of flat fields and polarization offset in the case of the ideal and real telescope guiding and pointing accuracy. The core algorithm in the formula is median process. Moreover, we obtained the flat fields and the crosstalk from Stokes I to Q , U and V using the above mentioned formula from series of the routine observed monochromatic image and Stokes Q / I , U / I and V / I images belonging to the Carrington Rotation 2144 observed by Huairou Solar Observing Station. The derived flat fields could not only reflect the small-scale dirty points but also the large-scale non-uniform pattern from the whole telescope system, which can be removed after the flat-fielding process. The distribution of the resulting Stokes V / I offset from Stokes I to V seems to be a Saddle surface, while that of the Stokes Q / I and U / I from I to Q , U is an upside-down bell shape. In order to evaluate the validity of the correction of the polarization offset we choose a quiet region near the solar limb as an example. The averaged value of Stokes V / I , Q / I and U / I in the region after polarization offset correction is −0.8×10−4, 0.9×10−4, 1.3×10−4, while that before correction is −6.8×10−4, 2×10−3, 1.7×10−3, respectively. It is quite clear that the Stokes V / I , Q / I and U / I improved a lot after the correction of the polarization offset. The main disadvantage of the method is that it can only correct the crosstalk from Stokes I to Q , U and V , incapable of the correction of that from Stokes V to Q and U . Some other methods are needed so as to make up the disadvantages. We suggest that the above mentioned method can be considered as one of the calibration methods for the space-borne or remote unattended full-disk solar observing instrument because it only uses the routine observation data without any other special calibration observation.