Since differential expression of microRNAs (miRNAs) has been found to be highly associated with several types of cancer, the goal of the present study was to identify an miRNA fingerprint as a non‑invasive diagnostic tool to detect urinary bladder cancer using the easily accessible samples of whole blood and urine. Blood and urine samples from 4 controls and from patients suffering from superficial and invasive bladder cancer were analyzed using miRNA microarray consisting of 754 human miRNAs from the Sanger database v14. Using RT‑qPCR technique, 6 of the differentially expressed miRNAs were validated in the controls (20 blood, 19 urine samples) and patients with superficial (18 blood, 16 urine samples) or invasive (20 blood and urine samples each) tumours. Three blood miRNAs (miR‑26b‑5p, miR‑144‑5p, miR‑374‑5p) were found to be significantly upregulated in invasive bladder tumour patients (P<0.05) when compared to the control group. The expression of 2 miRNAs (miR‑618, miR‑1255b‑5p) in the urine of patients with invasive tumours was significantly (P<0.05) increased in comparison to the control group. Blood miR‑26b‑5p detected the presence of invasive bladder tumours with 94% specificity and 65% sensitivity. The urine miR‑1255b‑5p reached 68% specificity and 85% sensitivity in the diagnosis of invasive tumours. This pilot study represents the first characterization of an miRNA profile for urinary bladder tumours in whole blood samples. In addition, it was shown that invasive bladder tumours could be identified by differentially expressed urine miRNAs. Further studies are needed to test the clinical usefulness for bladder cancer detection and surveillance.