The pursuit of microRNA target genes necessitates the creation of novel software and web services. MicroRNAs, abbreviated as short non-coding RNA molecules, hold a pivotal role in metabolic regulation, plant responses to environmental stress, and gene expression. Gaining insights into microRNA functions and investigating their target genes can advance drug development and address biotechnological challenges. However, the study and identification of microRNA targets within the genome present technical obstacles. MicroRNA molecules may not exhibit complete complementarity with their mRNA targets. These molecules either contribute to mRNA degradation or inhibit translation, and this process can transpire without full target complementarity. Consequently, the delineation of targets solely based on the principle of complementarity lacks unequivocal clarity. Moreover, a single microRNA molecule can correspond to multiple target genes simultaneously. The solution entails harnessing substantial datasets, employing machine learning techniques, and leveraging neural networks. In bioinformatics, neural networks serve a variety of functions, encompassing the analysis of biomedical data, diagnostics, prediction, classification, and nucleotide sequence segmentation. The pursuit and anticipation of microRNA targets through machine learning methods are currently undergoing vigorous development. A comparative assessment of contemporary neural networks for this task has been executed. A neural network-driven web service for microRNA prediction has been created. The server aspect of the service was developed using the Python programming language and the Flask library. The Mitar neural network, founded on deep learning, was employed. This network demonstrates heightened precision in predicting microRNA targets. We deliberate on the applications of miRNA prediction in gene expression analysis. Sustained research efforts are imperative to enhance the efficiency and broaden the capabilities of the developed computer system.