This study constructed a probability distribution model based on Copula to evaluate the probability of flood elements combinations, including peak and volume combination, volume and duration combination, and peak and duration combination. These combinations probability distribution maps and isograms were plotted. Comparison of the same return period flood designed from single elements and combination elements probability distribution was also made in this study. The results show that generalized GDP distribution fits flood peak and volume well, while exponent distribution fits flood duration suitably. Based on these two primary distributions, the Copula combination probability distribution fits flood elements combination appropriately, and generates a more secure design flood than traditional single element model. This study proves that the combination probability distribution model describes flood feature availably, and is meaningful to evaluate flood risk accurately.