The aim of the research is selection and genetic improvement of the source material of maize silica and sugar subspecies by methods of classical selection and experimental mutagenesis and synthesis on their basis of high-yielding, adaptive hybrids capable of stably realizing their yield potential in different growing conditions. Research methodology. Field experiments were conducted at the Cherkasy State Agricultural Research Station NSC "IZ NAAN", located in the central Forest-Steppe of Ukraine (Smila) during 2020-2021. The period of action of active temperatures (above 10 ° С) begins on April 20–28 and lasts 155–180 days. The probability of frost to -3 ° C during the sowing of corn (III decade of April and the first decade of May) is 10 and 5 %, respectively, and during ripening (September – first half of October) from 1 to 30 %. Sowing of the experimental material was carried out in the third decade of April in early May. During the growing season, phenological observations were made, during flowering the plants self-pollinated under parchment insulators, crossbreeding, evaluation and selection of the best samples were performed. Common field, laboratory and statistical research methods were used. Research results. Based on the morphobiological characteristics and assessment of the set of features of the samples of the sugar corn collection, the length of the period from emergence to flowering to 75 % of the panicles was determined, early-maturing samples were selected. When studying the genetic variability of the characteristics of the performance of inbred lines such as weight and length of the head, the number of rows of grains revealed the best lines, the average of which differs significantly from the standard. Based on genetic analysis of performance traits, it has been shown that the selection of starting material by cob weight and length may be most effective, as under favorable conditions these traits are controlled by additive genes, while other traits play a crucial role. Conclusions. The frequency of the obtained more productive self-pollinated lines on the original hybrid material with a significant excess of control indicators for such elements of productivity as cob weight, length and number of rows of grains is 7.19 ± 1.36 %. Selection based on "number of rows of grains" is effective, the best multi-row lines are selected based on genetic sources from Australia. As a result of the analysis of the sugar content in the grains, samples with a sugar content of 8.6% in the top sugar and 5.8 % in the white corn were selected. The coefficient of inheritance of quantitative traits, such as cob weight, length, number of rows of grains, number of grains in a row, is 0.71–0.93, which indicates the possibility of successful selection of these traits. All studied numbers exceed the hybrid standard by cob weight and number of rows of grains. Numbers 269, 272 and 276 and 277 were significantly better in terms of yield, with yields of 15.7–19.8 t/ha.