Ascomycetous fungi of the genus Fusarium, which was separated as a single taxonomic group in 1809 by Heinrich Friedrich Link, are distributed worldwide and have different ecological niches. They have been found in soil and plants as endophytes, saprotrophs and parasites. Infections of agricultural plants, caused by Fusarium fungi, lead to annual damage, estimated in hundred million dollars. In addition to decrease harvest quality and quantity, Fusarium fungi are able to produce toxic metabolites (fusariotoxins). The consumption of fusariotoxin-contaminated winter grain led to deaths of tens of thousands of people in Volga and Ural regions in 1930s-1940s. These factors make necessary investigation of morphology, biochemistry and genetics of those organisms. Nowadays there is no universal taxonomic system of the genus Fusarium, so it is not possible to identify an isolate to species level in many cases. The high variability of morphological structures, and, on the other side, their similarity for closely related species, are major problems for researchers which use classical methods for identification and systematization of this group of fungi. Today, the molecular methods, based on the use of specific DNA sequence, are playing an increasingly important role in Fusarium systematics. The application of this approach has led to establishment of a number of new species in the genus. T. Yli-Mattila et al. (2009, 2011) used multilocus phylogenetic analysis to describe the novel species F. sibiricum and F. ussurianum which were found in Siberia and Far East and proved to be morphologically similar to F. graminearum and F. langsethiae, respectively. The authors of this paper identified the species F. torulosum (morphologically similar to F. avenaceum) in Russia for the first time based on the use of DNA markers. It is no less important that the analysis of inter- and intraspecific polymorphism makes possible the development of highly-specific assays for molecular detection of Fusarium fungi, including ones which infect plants and produce mycotoxins. The use of these assays (for example, ones described by M. Nicolaisen et al., 2009 and A.A. Stakheev et al., 2011) allows not only to detect the contamination of plants by mycotoxin producers, but also to estimate the quantity of their content in a probe. At the same time, a lot of questions are still unanswered: particularly, simultaneous application of different species concepts (morphological, biological and phylogenetic) makes it difficult to develop universal taxonomic system of the genus; also the most informative and reliable marker, which would reflect the evolutionary history of the genus, has not been found. The main achievements and problems of Fusarium phylogenetics, and perspectives of the application of molecular approach for studies of this group of organisms are discussed in this review.