The invention relates to the field of hydropower scheduling, and relates to a method for multi-objective optimal operations of cascade hydropower plants based on relative target proximity and marginal analysis principle. The invention constructs an optimization model with relative target proximity by introducing positive and negative ideal points of objectives so that a multi-objective optimization is efficiently handled. A strategy for processing complex operation constraints is proposed and coupled with genetic algorithm to solve the model. Thus, optimal solution sets under different weight coefficients of objectives can be determined. Following this, marginal benefit and marginal cost are introduced to represent a relationship between multiple objective values and corresponding weight coefficients. According to principle of profit maximization in economics, a marginal analysis of objective weight coefficient is finally made to determine benefit-dominant area, cost-dominant area and equilibrium area, and there by obtaining a basis of multi-objective decision-making.