Abstract: The short-term scheduling of gradient hydropower needs to guarantee the ecological requirements of the downstream river section of each power station, and the in-depth study of ecological scheduling requires that ecological constraints become more and more diverse, which makes the short-term optimal scheduling problem of basin gradient more and more complicated. In this paper, the ecological constraints are firstly classified into Class Ⅰ, Class Ⅱ, and Class Ⅲ according to the flow constraints and control modes to meet the complex ecological scheduling requirements of the gradient reservoirs; then, the short-term day-ahead peaking model of gradient hydropower coupled with complex ecological constraints is constructed to minimize the peak-to-valley difference of the residual loads; finally, the segmented linearization technique is used to deal with the one-dimensional hydraulic nonlinear constraints, and the quadrilateral mesh gridded technology is used to deal with the binary power nonlinear constraints by combining the special ordered sets of type 2(SOS2) constraint modeling method, and the Big M method is used to linearize the more complex class Ⅲ ecological constraints, then transform the original model into a standard mixed-integer linear programming (MILP) model for solving. Taking the six cascade reservoirs in the mainstream of Wujiang River (Guizhou section) as the engineering background, the validity of the model is verified under different water inflow scenarios in the flood and dry periods, and the computational results are significantly superior to the dynamic programming successive approximation (DPSA) algorithm, through the comparative analysis of the results with the three reference models, it is concluded that the setup and type of ecological constraints will affect the peaking effect, and the necessity of reasonable design of the ecological flow of water conservancy projects is emphasized.