PURPOSE: To evaluate the predictability of amount of hemothorax in the patients with blunt chest trauma, supine chest AP radiographs of 66 patients were reviewed and statistically analyzed. MATERIALS AND METHODS: In 66 patients, rib fractures were present in 53 patients, hemothorax in 46 patients, pneumothorax in 25 patients, and pulmonary contusions in 18 patients. Width and length of hemothorax were measured on supine chest AP radiograph, and were correlated with known drained amount of hemothorax. The presence and number of rib fracture, pulmonary contusion, subcutaneous emphysema, fracture of scapula and clavicle, and total opacification of hemithorax were also correlated with the drained amount of hemothorax. RESULTS: In multiple logistic regression analysis, width of hemothorax had the highest correlation with drained amount of hemothorax(regression coefficient 0.718, p value 0.00005). The presence and number of rib fracture, scapular fracture, subcutaneous emphysema were also correlated with drained amount of hemothorax. But length of hemothorax, pulmonary contusion, clavicular fracture, total opacification of hemithorax were not correlated with drained amount of hemothorax. CONCLUSION: Measured width of hemothorax in supine chest AP radiograph is the most reliable predictor for estimation of the amount of hemothorax, and may also be used as an indication for the application of closed thoracostomy in the treatment of hemothorax.