A new data evaluation method and integrated software have been developed for the quantification of individual aerosol particles based on an iterative reverse Monte Carlo simulation combined with successive approximation for the elemental composition. The computer code supports the automatic spectrum processing and the statistical analysis by clustering of the measured and calculated data of the particles: X-ray characteristic intensities, calculated elemental concentrations and the particle sizes. The analytical procedure was tested rigorously by measurement of standard single particles such as (NH₄)₂SO₄, NH₄NO₃, CaCO₃, CaSO₄, SiO₂, Fe₂O₃, BaSO₄, KNO₃, NaCl and a good agreement between the nominal and calculated quantitative composition was found within 5–15 relative %. The correction of the second order processes caused by the characteristic fluorescence line of the substrate material (AlKα) on the calculated concentration was estimated theoretically by a single mathematical model for particles with a spherical shape. The k ratio for the fluorescence correction was found to be less than 0.1–0.7% for low-Z analysis. The present semi-automated method was applied to analyse marine aerosol samples collected over the North Sea. Results of approximately 500 small individual particles show the capability of the method to quantify the elemental composition of sub-micrometre particles down to 0.2 μm.