Exposure to ionizing radiation from nuclear devices, spaceflights or terrorist attacks represents a major threat to human health and public security. After a radiological incident, noninvasive biomarkers that can facilitate rapid assessment of exposure risk in the early stages are urgently needed for optimal medical treatment. Serum microRNAs (miRNAs) are ideal biomarkers because they are stable in response to environmental changes, they are common among different species and are easily collected. Here, we performed miRNA PCR arrays to analyze miRNA expression profiles at 24 h postirradiation. Blood samples were collected from animals that received 0.5–2 Gy total-body carbon-ion irradiation. A specific signature with 12 radiosensitive miRNAs was selected for further validation. After exposure to 0.1–2 Gy of carbon-ion, iron-ion or X-ray radiations, five miRNAs that showed a significant response to these radiation types were selected for further observation of dose- and time-dependent changes: miR-183-5p, miR-9-3p, miR-200b-5p, miR-342-3p and miR-574-5p. We developed a universal model using these five miRNAs to predict the degree of exposure to different radiation types with high sensitivity and specificity. In conclusion, we have identified a set of miRNAs that are quite sensitive to different radiation types in the early stages after exposure, demonstrating their potential use as effective indicators to predict the degree of exposure.