The performances of four chromatographic methods in combination with multivariate statistical data analysis were compared for the determination of cocoa butter equivalents (CBEs) in mixtures with cocoa butter (CB). The sample set analysed consisted of 42 different CBs, including specific crops (soft South American and hard Asian CB types), and 21 commercially available CBEs, including several types of illip-containing fats. From these pure samples about 250CB/CBE mixtures were prepared containing between 5% and 20% CBE addition. The triglyceride (TG) composition of the blends was analysed by packed column gas-liquid chromatography (GLC), capillary GLC and high-performance liquid chromatography. In addition, the fatty acid profile was determined by capillary GLC. Data generated by packed-column GLC were evaluated by a graphical procedure, whereas multiple linear regression analysis was applied to data from high-resolution methods. The calibration model that most efficiently minimised the error of prediction was based on a combination of fatty acid and TG data as input variables. The average error of prediction was estimated to be 2.3% CBE in CB, without prior knowledge of either the CB or the CBE composition. Recalculated for a chocolate with a fat content of 20% the error estimate would equal 0.5% CBE in the finished product.