Aiming at the urban waterlogging problem induced by urban rainstorm and outer river flood, the characteristics of urban rain-flood encounter were studied. Taking Duanzhou District of Zhaoqing City, Guangdong Province as an example, rain-flood joint distributions were constructed with symmetric and asymmetric product Copula functions, respectively, based on long series of measured rainfall data and Xijiang River streamflow data. The optimal Copula function was selected by Akaike information criterion and the graph evaluation method for rain-flood encounter risk analysis. The results show that the asymmetric product Frank Copula II function has the best fitting effect. In the rainstorm-dominated combination, the conditional risk probability of Xijiang River flood is relatively high when heavy rain occurs in Duanzhou District. In the flood-dominated combination, the rain-flood combined risk probability under different flood recurrence periods is slightly greater than the corresponding frequency of Xijiang River flood.