The impact of long-term continuous cropping on the Fusarium population abundance and diversity, pathogenicity and phylogeny in soybean field were analyzed by using isolation, morphological identification, pathogenicity test, sequencing analysis and molecular marker with restricted fragment length polymorphisms (RFLP). The soybean field was located at the Hailun Experimental Station of Agricultural Ecology of Chinese Academy of Sciences in Northeast China and had been under a long-term rotation experiment designed to two treatments, i. e., long-term continuous cropping (LCC) of soybean for 20 years and short-term continuous cropping (SCC) for 3 years. In SCC field, the population density of Fusarium spp. was 6.0 x 10(4) CFU x g(-1), in which F. oxysporum, F. graminearum and F. verticillioides possessing high pathogenicity and F. solani possessing moderate pathogenicity were the dominant species. In LCC field, the population density of Fusarium population and the dominance index of dominant species were significantly lower than that in SCC. The population density of F. oxysporum, F. graminearum and F. solani were only 36% , 32% and 22% of that in SCC, and F. verticillioide with highest pathogenicity was absent. The diversity and evenness index of Fusarium population were significantly higher than that in SCC. F. tricinctum, F. lateritium and F. avenaceum, just isolated from LCC, possessing a distant genetic relationship with Fusarium isolates possessing high pathogenicity based on internal transcribed spacer (ITS) and translation elongation factor 1-alpha (EF-1alpha) gene, were non-pathogenicity for soybean. Thus, it seemed that LCC of soybean could cause the inhibition of soil Fusarium population size, alteration of Fusarium community composition and genetic diversity, and even the decline of pathogenicity for soybean root rot disease of Fusarium population.