Characterization of stratospheric liquid ternary solution (LTS) aerosol is studied as an inverse problem in retrieving particle volume size distribution and chemical composition of ternary droplets using the extinction coefficient over wavelengths of 3 to 12 ¿m. The retrieval method comprises a constrained, weighted least squares minimization procedure between measured and modeled spectra. The constraints restrict the first derivative of the particle size distribution with respect to particle size and restrict the relationship between weight percents of discrete partials (H2SO4/H2O/HNO3) in the ternary droplets using the Carslaw equilibrium model. In the inverse modeling, emphasis is given to minimize the number of reference spectra up to consideration of two mostly binary components composed of H2SO4/H2O and HNO3/H2O solutions. This permits the retrieval of particle volume size distribution independently of particle chemical composition and therefore simply adopts the inversion method for the processing of spaceborne occultation measurements when other aerosol/cloud components along with trace gases are subject to identification. We initially carry out numerical tests by applying the inversion method to synthetic extinction measurements in order to study the limitations of the retrievals over wide ranges of microphysical parameters. A case study in processing actual space-borne data obtained by the Improved Limb Atmospheric Spectrometer (ILAS) demonstrates the efficiency of the retrievals.