Diabetic foot ulcer (DFU) is the most serious and costly chronic complication that may lead to disability and even death in patients suffering from diabetes mellitus (DM). However, the clinical diagnosis and prognosis of DFU is inadequate. There is still a lack of effective biomarkers for its early diagnosis. We obtained the circRNA expression dataset GSE114248 and mRNA expression dataset GSE80178 from the GEO. R software was used to identify the differentially expressed circRNAs (DECs). The mRNAs associated with DFU were identified by a random forest algorithm and intersected with mRNAs predicted by circRNAs. Then, the circRNA-miRNA-mRNA network was established and the hub genes were screened using GO semantic similarity and were validated by the GSE199939 dataset. Meanwhile, the expression level of the biomarkers was verified by RT-PCR assays and immunohistochemistry. Finally, GSEA was conducted to determine differential immune cell infiltration and the immunological cells' relationships with hub genes. We identified three hub genes including KIAA1109, ENPP5, and NRP1 that might play an important role in DFU. ROC curve results also showed a good performance of these three genes in the validation dataset. Furthermore, RT-PCR assays and immunohistochemistry confirmed the results above. Immune infiltration analysis indicated that DFU had a significant increase in Neutrophils. Moreover, three hub genes were closely correlated with a variety of inflammatory cells. KIAA1109, ENPP5, and NRP1 are key hub genes of DFU. They might play an important role in the development of DFU and could be potential biomarkers in DFU.