Background and Aims Intrahepatic cholangiocarcinoma (ICC) is a malignant tumor arising from the intrahepatic bile duct epithelium, which has an insidious onset and high degree of malignancy. Because there are no obvious clinical symptoms in the early stage of ICC, and most patients have lost the opportunity for surgery at the time of diagnosis, so its prognosis is very poor. Exploration of targets for early diagnosis and treatment of ICC is of great significance. Therefore, this study was conducted to screen the key genes involved in the occurrence and development of ICC.Methods Two ICC transcriptome datasets (GSE107943, GSE119336) were downloaded from the GEO database. The differentially expressed genes were screened by edgeR package of R language, and then these genes underwent GO and KEGG pathway enrichment analysis. The protein-protein interaction (PPI) networks of these genes was constructed by using STRING database, and the key protein regulatory genes were mined by using the MCODE plug-in of Cytoscape. The expressions of key protein regulatory genes in tumor tissues were analyzed and verified by UALCAN and GEPIA databases. UCSC XENA database was used to analyze the expressions of key regulatory genes in generalized carcinoma. TCGA database was used to analyze the co-expression genes of the key regulatory genes. UALCAN and GEPIA databases were used to analyze the relationship between key regulatory genes and patient prognosis, tumor grade, stage and lymphatic metastasis. The correlation between the expressions of the regulatory genes and immune infiltration were calculated using R language GSVA package. The predictive abilities of the key protein regulatory genes to ICC were evaluated by drawing ROC curve. Cell experiments were performed to verify the expressions of the key regulatory genes.Results A total of 1 094 common differentially expressed genes were screened, including 567 up-regulated genes and 527 down-regulated genes, which were mainly involved in the process of small molecule catabolism, organic acid biosynthesis, carbon metabolism and so on. Three key genes Polo-like kinase 1（PLK1）, hydroxyacid oxidase 2（HAO2）and ficolin-2（FCN2） were mined through PPI networks. PLK1 gene was significantly up-regulated in tumor tissues, and HAO2 and FCN2 genes were significantly down-regulated in tumor tissues, which were verified by UALCAN and GEPIA databases. The analysis of UCSC XENA database showed that the expression of PLK1 was significantly increased in 28 types of tumors, the expression of HAO2 was significantly decreased in 24 types of tumors, and the expression of FCN2 was significantly decreased in 27 types of tumors. The analysis of TCGA database showed that PLK1 was co-expressed with CCNA2 and GTSE1, HAO2 was co-expressed with MTTP and CPS1, and FCN2 was co-expressed with FAM99A and GDF2. The analysis of the UALCAN database found that the expression of three genes was related to the stage and grade of the tumor and lymph node metastasis. Among them, high expression of PLK1, low expression of HAO2 and FCN2 were associated with higher tumor stage, worse differentiation and more prone to lymph node metastasis. The correlation analysis found that the expression of PLK1 was significantly positively correlated with the infiltration of Th2 cells, and the expression of FCN2 was significantly negatively correlated with the infiltration of aDC cells. The ROC curve showed that all the three genes could diagnose ICC well, among which HAO2 had the best diagnostic ability. The results of cell experiments showed that the expression of PLK1 was significantly increased, while the expression of HAO2 and FCN2 were significantly decreased in RBE （all P<0.01）.Conclusion LK1, HAO2 and FCN2 may be the key protein regulatory genes involved in the occurrence and progression of ICC. These three genes may probably become new targets for the diagnosis and treatment of ICC.