Clear cell renal cell carcinoma (ccRCC) is a heterogenous tumor with high metastatic potential. Circular RNAs (circRNAs) play key roles in cancer initiation and progression. However, the knowledge of circRNA in ccRCC metastasis is still inadequate. In this study, a series of in silico analyses and experimental validation were employed. The differentially expressed circRNAs (DECs) between ccRCC and normal or metastatic ccRCC tissues were screened out using GEO2R. Hsa_circ_0037858 was identified as the most potential circRNA related to ccRCC metastasis, which was significantly downregulated in ccRCC compared with normal and was also markedly decreased in metastatic ccRCC compared with primary ccRCC. The structural pattern of hsa_circ_0037858 presented several microRNA response elements and four binding miRNAs of hsa_circ_0037858, consisting of miR-3064-5p, miR-6504-5p, miR-345-5p and miR-5000-3p, were predicted using CSCD and starBase. Among them, miR-5000-3p with high expression and statistical diagnostic value was considered as the most potential binding miRNA of hsa_circ_0037858. Then, protein-protein interaction analysis revealed a close linkage among the target genes of miR-5000-3p and the top 20 hub genes among them were identified. Based on node degree, MYC, RHOA, NCL, FMR1 and AGO1 were ranked as the top 5 hub genes. FMR1 was identified as the most potential downstream gene of hsa_circ_0037858/miR-5000-3p axis according to expression, prognosis and correlation analysis. Moreover, hsa_circ_0037858 suppressed in vitro metastasis and enhanced FMR1 expression in ccRCC, which could be markedly reversed by introduction of miR-5000-3p overexpression. Collectively, we elucidated a potential hsa_circ_0037858/miR-5000-3p/FMR1 axis involved in ccRCC metastasis.