Facing on the bad convergence and easy premature often emerged in Genetic Algorithm making solution of hydropower station economic load dispatch, the Breed-evolutionary Improved Genetic Algorithm (BIGA) was proposed. BIGA used the partial solution constraint initial population generation method to avoid cavitations-vibration range, so as to describe the level of group-evolution, and define the function of population -diversity and threshold of population -diversity. Meanwhile, the global searching ability of Genetic Algorithm and local searching ability of Breed were effectively applied. A contrast between BIGA and SGA applied for the Three Gorges hydropower station was presented. The distribution results based on several loads show that: In BIGA, stable and safe operation of unit were ensured by avoiding cavitations-vibration range. And, with the strong local searching ability of breed, the population diversity was maintained, and the searching ability and convergence of the algorithm were improved.