Abstract.A calibration model has been developed in order to overcome matrix effects in atomic absorption measurements. The model uses two independent variables for analyte quantification (the amount of the sample and the amount of analyte added). The dependent variable is the absorbance measured. The method also allows matrix interferences to be controlled without prior knowledge of matrix composition. The method is applied to iron determination by FAAS in the presence of large amounts of copper. Direct calibration and standard addition are also performed in order to compare them with the new empirical model. Results show that the error in iron determination could be -42% when direct calibration is applied and + 10% when the standard addition method is used, whereas the proposed model decreases the error to -20%.