Abstract: This paper presents a method to filter and replace the erroneous extreme data in historical hydrological series by data preprocessing and screening to improve the accuracy of monthly frequency calculations for instream ecological flow. We formulate the problem from the economic point of view and treat the ecological flow as a penalty factor in protection of the ecological environment of rural cascaded hydropower stations. This is different from the previous studies that take the ecological flow as a constraint or one of the multi-objectives in an optimal scheduling model. We have developed a new general model that is solved by DNMPSO, and applied it to the Lushui River cascaded stations in Jiangxi. The results show that this new model gives schemes that can better protect the downstream through restoration of ecological environment and also improve the overall benefits of the cascaded stations.