In order to solve the inadequacies of existing calculation and allocation methods of the power generation compensation benefits (PGCB) which come from united operation of cascade reservoirs, a new concept, electric energy benchmark of hydro plant, was proposed and used to calculate PGCB, and also a new model for allocating PGCB among cascade hydro plants was developed, which can balance the subjective preference of each plant and the group consistency of all plants. The proposed computing method takes into account TOU power price differences among different plants and load rate constraint from power grid for the electric energy output increasing of every plant, which makes the method get closer to the real value of PGCB. According to the benefits-sharing principles selected by involved parties, the proposed model firstly gets the individual preference weight vector of every plant for the principles by solving an optimal model whose goal is maximum single plant apportionment ratio; secondly, reach a consensus on the weight vector through maximizing group consistency; finally, the actual benefits –sharing proportion of every plant is fixed by solving another optimal model whose goal is minimizing the distance between the final values and the results from all the different principles. Moreover, an improved partheno genetic algorithm, whose genetic operation is simple and more efficient, was presented and adopted to solve all the relevant optimal models. A case study shows that the proposed computing method and benefits-sharing model work well, and the results are reasonable and can be accepted more easily.