Аннотация:This paper considers methods developed by the author for thematic search of scientific and technical journals and conferences, using software tools that implement full-text search and expert search algorithms. First, a brief review of the existing systems of thematic search for scientometric data is given with an indication of their main shortcomings. Algorithms developed by the author, their software implementation, and testing on big data of the scientometric system are described. To execute queries, the algorithms developed by the author use full-text descriptions of objects in scientometric systems and graphs of mutual relations between objects, filter data by narrowly focused thematic queries, and rank search results by authority.