The paper focuses on the identification of geographical origin of fruits and vegetables based on their composition using statistical and chemometric methods. 58 samples of potatoes and 22 samples of apples from various districts of Kharkiv city and Kharkiv region (Ukraine) were analyzed. Concentrations of Fe, Mn, Zn, Cu, Ni, Pb, Al, Co, Cr, and Cd ions in the samples were determined with the use of atomic absorption spectroscopy. Arrays of analytical results were processed with the use of different statistical and chemometric methods, including calculation of the Spearman's rank correlation coefficients, Kruskal-Wallis and Wilcoxon-Mann-Whitney tests, Principal Component Analysis. Probabilistic neural network were shown to be the efficient tool for identification of geographical origin of potatoes and apples samples.