Fusarium graminearum causes Fusarium head blight, damaging small-grain crops including barley, and produces the mycotoxin deoxynivalenol (DON), which poses health risks to humans and animals. Crop rotation is a sustainable management strategy often used in North Dakota to control various pests and pathogens. However, crop residues can harbor pathogens such as F. graminearum, serving as inoculum sources for subsequent crops. Although F. graminearum is mainly a fungal pathogen of cereals, it has also been reported to cause dry rot in potatoes and root rot in soybeans. The pathogen adapts to different crops, which makes crop rotation alone insufficient for managing the disease. We hypothesize that, in addition to its preference for cereals, F. graminearum might also favor non-cereal crops in a rotation system. To examine this, we cross-infected a total of seven F. graminearum isolates collected from barley (2), potato (1), soybean (3), and winter rye (1) onto barley, potato, and soybean. Results of our cross-infectivity experiments show that F. graminearum isolates vary in aggressiveness levels. For instance, isolates from potato and barley caused similar levels of disease when inoculated into potato tubers, whereas soybean and winter rye isolates caused no to low disease in potato. We sequenced isolates and examined how genomic features affect host preferences. Each genome had 4–5 contigs totaling 36.5-37.5 Mb, consistent with F. graminearum size. The largest contigs reached 11.8 Mb, with an N50 over 9 Mb, indicating high contiguity. GC content was stable at 47.8%, with no ambiguous bases. Genomic analysis showed that F. graminearum isolates from barley and potato had higher variant counts and rates than those from soybean and winter rye. In addition, barley and potato had the most SNPs, insertions, and deletions, while soybean and winter rye had fewer variants. Variant rates were also higher in barley and potato (1 per 233–236 bp) than in other isolates (1 per 337–390 bp). All seven isolates share a large core genome of 11,505 orthogroups, indicating conserved functions, but comparing accessory genes reveals two groups: barley and potato, and soybean and winter rye. Our findings indicate that while F. graminearum has conserved regions (core), a small part of its genome (accessory) is variable and may influence the fungal pathogen’s ability to infect specific crops. Our study is the first to demonstrate the connection between cross-infectivity patterns and genomic variation in Fusarium graminearum.