Objective：To find more differentially expressed circRNAs in thyroid cancer and explore their mechanisms through microarray analysis. Methods：R software was used to obtain differential expressed circRNAs from the microarray datasets GSE93522. CircRNA-miRNA-mRNA regulatory networks were constructed by Cytoscape software according to the relevant miRNAs and mRNAs predicted by miRDB，miRTarBase and TargetScan，combining with differential expressed miRNAs and mRNAs from The Cancer Genome Atlas（TCGA）. The important circRNA-miRNA-mRNA networks were obtained by using gene ontology（GO）analysis，Kyoto encyclopedia of genes and genomes（KEGG）analysis，survival curve analysis and so on. Furthermore，the vital regulatory networks were verified by GSE40807，GSE3467，GSE33630 and clinical data from TCGA. Results：The circRNA-miRNA-mRNA regulatory networks included 18 circRNAs，8 miRNAs and 26 mRNAs. GO analysis showed that some of the mRNAs were involved in cell metabolism，migration，and other functions，as well as the composition of cyclin-dependent protein kinase complex. KEGG analysis showed that some mRNAs were correlated with the PI3K-AKT pathway，MAPK pathway，and P53 signaling pathway. Important circRNA-miRNA-mRNA regulatory networks were constructed according to the circRNA score and function analysis，which based on 6 circRNAs，8 miRNAs and 16 mRNAs. Hsa_circ_0001658-hsa-mir-183-AKAP12 was considered to the vital regulatory network according to the expression verification and clinical data analysis from TCGA. Conclusion：Many circRNA-miRNA-mRNA networks are related to the development of thyroid cancer. Hsa_circ_0001658-hsa-mir-183-AKAP12 is considered to a vital regulatory network. These findings may be helpful to find more diagnostic or prognostic markers and therapeutic targets of thyroid cancer.