The continuous raise of molecular diversity requires to find a guidelines for orientation it this. A likely occasion for consideration of new QSAR approach possibilities for this aspect is the flavonoids (F) multitude, the class of organic compounds (C) with very varied chemistry and pharmacology. The list of F affinities includes benzodiazepines, adenosine, estrogen receptors and row of enzymes - phosphodiesterase, proteinthyrosine kinase, aromatase, xanthine oxidase asf. The simple way of clasterization of C is their numerical representation as the vector in linear space. The components of the some vectors are determined by parameters of molecular shape (number of cycles, tertiary atoms asf) and orbital structure indices (numbers of occupied orbitals of different kind). The nonesterificied F are tricyclic C with numbers of tertiary atoms n = 5,...,11 - and more, from generic flavon (flavan), n = 5, to myricetin (n = 11) and other. The all derivatives of F with given n may be associated with vectors of linear space, sgr;-set, with linear form s = 2l + m, l - number of nonhydrogenous atoms, m = 0, ±1,... The collection of σ-sets which numbered by given number n = 5,...11,... envelops all natural F and their derivatives which have not additional cycles. In mentioned σ-sets of tricyclic C by the some mode are presented ligands of benzodiazepines, adenosine, estrogens and other receptors and enzymes. Positions of ligands of each class form a triangle ”sector of affinity“ in 4-5 adjacent σ-sets. These triangle sectors in mentioned collection of σ- sets are disposed in conformity to certain laws. The position of given F relatively to ”sectors of affinity“ allow to found their affinity (and cross-affinity) to different targets. Calculations with utilization of sophisticated shape parameters and electronic structure indices allow to determine the affinity and / or activity quantitative measure.