Background: CRC has a high mortality rate, and early detection is essential to reducing the CRC-related mortality rate. Screening is inefficient for diagnosis based on the low detection rate, length of time consumed, and poor compliance, so more sensitive and effective diagnostic markers are needed to improve diagnostic efficiency. Aldo-keto reductase family-1 enzymes (AKR1s) play a crucial role in the NADPH-dependent reduction of various carbonyl substrates (aldehyde, ketone), and their expression is closely linked with the progression and invasion of malignant cancers. However, the expression patterns of AKR1 family members in colorectal cancer (CRC) and their unique roles in the diagnosis of CRC have not been explored. Here, the mRNA and protein expression levels of 10 AKR1s and their unique roles in diagnosis in CRC were clarified. Methods: We examined the transcription, methylation and survival data of AKR1s in CRC patients from the Oncomine, Gene Expression Profile Interactive Analysis (GEPIA), Kaplan‒Meier Plotter, cBioPortal and Illumina databases. Then, the mRNA and protein expression levels in CRC tissues were assessed by qRT‒PCR and Western blotting. Results: This study showed that the transcript and protein expression levels of AKR1B1, AKR1B10, AKR1B15, AKR1C1, and AKR1C2 were decreased in CRC tissues compared with adjacentnormal tissues. In contrast, the expression levels of AKR1A1 and AKR1C4 were increased in CRC tissues compared with normal tissues. These expression trends were mostly consistent with those for the Oncomine, GEPIA, TCGA and cBioPortal database analyses, and AKR1B1, AKR1C1 and AKR1C2 were significantly downregulated based in their high CpG methylation levels. The qRT–PCR analysis of AKR1 expression in CRC patients with different clinicopathological characteristics showed that the AKR1A1 expression level was related to invasion depth, and the AKR1C2 expression level was significantly related to TNM stage and distant metastasis. The GEPIA results showed that the malignant degree increased with the increase of AKR1C1, AKR1C2 and AKR1C4 concentration, but the opposite was true for AKR1A1, and ROC curve analysis showed that a low level of AKR1B10 and a high level of AKR1C4 had high diagnostic efficacy in the diagnosis of CRC, with sensitivity values of 71.8% and 70.4% and specificity values of 93.0% and 79.6%, respectively. Conclusion: In summary, the study revealed that AKR1A1 and AKR1C4 have good diagnostic value for CRC and could become potential tumour biomarkers. AKR1B1, AKR1B10 and AKR1B15 may be favourable prognostic indicators.