Abstract: Flood coincidence probability analysis among different regions is quite important for flood design and controt.The current flood coincidence probability analysis is based on the simple statistical analysis of observed data without theoretic methods.In this study,the daily flow data of four hydrological stations in the upstream Yangtze River and its tributaries,including the Pingshan at the Jinsha River,Caochang at the Min River,Beibei at the,lialing River and Yichang at the Yangtze River,are selected for case study.Multidimensional copula functions are introduced and used to construct the joint distributions of flood occurrence dates and magnitudes.Flood coincidence and conditional probabilities among the upstream Yangtze River and its tributaries thus are estimated.The calculated flood encounter values are compared to the empirical ones estimated by the observed data.The results show that the proposed method is rational and provides a new approach for flood encounter analysis.