Abstract: Objective:To evaluate the value of using fluorescence in situ hybridization (FISH) for detecting the chromo -some aberration of urine exfoliated cells in urothelial carcinoma. Methods:The DNA probe specific for the centromere re -gion of chromosomes 3, 7 and 17and for p 16site was used to detect the FISH in the urine of 20healthy subjects to set up the threshold. FISH detection of the urine was then conducted on 115 patients who were suspected of blood urine from uro-thelial tumor, with the positive diagnosis being declared if the result of detection using at least 2 probes was over the thresh-old or there were at least 2 abnormalities in the detection using 1 probe. At the same time, conventional H&E staining was used for a morphologic analysis of the cast-off cells in the urine. Results: The sensitivity of the FISH detection and of the di-agnosis of exfoliative cytology for urothelial carcinoma were respectively86.7% (78/90) and 10% (9/90), (P<0.001 ), and the specificity of the detection and diagnosis were respectively 96% (24/25) and 100 % (25/25), (P>0.05). The positive predic -tive value was 98.7% (78/79) vs. 100 % (9/9), (P>0.05), and the negative predictive value was 66.7% (24/36) vs. 23.6% (25/106 ), (P<0.05). The sensitivity of FISH technique for detecting urothelial carcinoma has no correlation with the tumor grad -ing and staging. The positive rates of the diagnosis in the low and high-grade urothelial carcinoma were 85.7% and 87.5%, respectively, (P>0.05). The positive rate in pTa-1 and pT2 –4 were84.2% and 88.4% , respectively (P>0.05). Conclusion :FISH technique is a non-invasive and effective method for the diagnosis of urothelial carcinoma and is more sensitive and specific than urinary cytology. FISH technique can be a useful tool in the early detection of urothelial carcinoma and could also be of important clinical significance in predicting the biological behavior and prognosis of a carcinoma.