The feasibility to discriminate among samples of different fat blends prior and after inorganic or lipase-catalyzed interesterification, via pattern recognition techniques [principal component analysis (PCA) and discriminant analysis (DA)], was investigated. Blends I and II, consisting of mixtures of palm stearin, palm kernel oil and a concentrate of triacylglycerols (TAG) rich in n-3 polyunsaturated fatty acids (EPAX 4510TG or EPAX 2050TG) were used. These blends, prior (64 samples) and after interesterification, catalyzed by an immobilized Thermomyces lanuginosa lipase (Lipozyme TL IM, 54 samples) or by sodium methoxide (10 samples), were characterized by their acylglycerol profiles (25 chromatographic peaks) and solid fat content (SFC) at 10, 20, 30 and 35 °C. PCA on the multivariate data (i) showed that the initial samples were characterized by higher SFC and higher contents of high-melting TAG and (ii) suggested two separate clusters of initial and interesterified samples. DA was performed on the multivariate data to determine which of the 29 variables have discriminative power. When the 124 samples, characterized by their acylglycerols, were grouped into (i) initial and interesterified samples of blends I or II (four groups) or (ii) also by the catalyst used (six groups), 98.4% of the samples were correctly classified.