Purpose. To develop a tobacco variety model with optimal inflorescence traits such as size and shape that will allow to increase seed productivity of the crop. Methods. Statistical and mathematical (correlative, regressive) ones. Results. Basic collection consisting of 282 variety samples registered in the National Genetics Center was evaluated, optimal parameters of inflorescence were defined that can provide a high seed yield. During statistical analysis, correlation matrix was developed with the purpose to highlight traits that correlate with inflorescence productivity. According to the results of correlation analysis, a strong relationship between the width and height of inflorescence (r = 0,773±0,038) was established. Somewhat weaker correlation was observed when modeling regressive relation between inflorescence height and width, where regression showed the medium relationship. Regression equation of these traits is as follows: y = 0,5585x + 8,4649. Inflorescence density (r = 0,646), height (r = 0,556) and width (r = 0,527) also had quite a high positive effect on seed productivity. The results of regression analysis pointed to the fact that there were a linear relationship between inflorescence size and seed productivity. Conclusions. Among 282 samples of basic tobacco collection, 29 varieties with high seed productivity was defined which can be used in the breeding process, and ‘Sobolchskyi 15/21’, ‘Ergo 23’, ‘C-11’, ‘Sygarnyi 99’ varieties were selected for large-scale implementation into the production of raw material of cigar type.