Improving　the　calibration　accuracy　of　infrared　hyperspectral　interferometers　is　a　prerequisite　for　their　quantitative　application.　Updating　satellite　in-orbit　calibration　parameters　is　an　important　way　to　keep　calibration　accuracy.　However,　how　to　separate　the　deviation　of　one　or　more　calibration　parameters　from　the　coupling　error　of　observation　data　is　a　key　issue　that　needs　to　be　solved.　Therefore,　we　propose　a　variational-based　calibration　parameter　optimization　algorithm　(VarCalPOA),　which　uses　only　observation　data　and　reference　data　to　optimize　the　key　calibration　parameters　based　on　variational　assimilation　theory　to　obtain　the　optimized　values　of　the　calibration　parameters,　thereby　improving　the　calibration　accuracy.　Based　on　the　observation　and　calibration　simulation　model,　we　choose　T-ict,　e(ict),　and　a(2)　as　the　key　calibration　parameters.　Using　the　VarCalPOA　method,　the　simulation　results　show　that　the　optimized　calibration　parameters　have　a　certain　positive　effect　on　the　improvement　of　calibration　accuracy.　This　study　proves　that　the　VarCalPOA　method　is　potentially　an　effective　method　for　monitoring　the　in-orbit　state　of　infrared　hyperspectral　interferometers.