Abstract: Exploring new solving methods has always been one of the hotspots for the study on optimal reservoir operation. The social emotional optimization algorithm (SEOA) is a new heuristic intelligent optimization algorithm, but it has not been applied to reservoir operation yet. In application of SEOA to the operation of cascade hydropower stations, this study formulates an initial population uniform design to overcome the low representativeness of initial population caused by stochastic generation, and develops a population elimination strategy aiming at the low activity of population and the tendency of local convergence caused by some individuals’ premature convergence. Thus an improved social emotional optimization algorithm (improved SEOA) is achieved. A case study shows that this algorithm is efficient and stable, and it has a great search capability in solving the optimal operation models of cascade hydropower stations.