Improved Particle Swarm Optimization (PSO) is applied to solving the short-term optimal scheduling of Songjianghe cascade hydropower stations. Two mathematic models characterized respectively by maximum  power generation and maximum  generation profit are established. Shrinkage factor and hereditary mutation operator are adopted to overcome the shortcomings of standard PSO like precocity and slow convergence in late stage. These  mathematical models are employed in the optimal scheduling analysis for Songjianghe cascade power stations, and the results show that the quality and benefit of the stations could be improved by short-term optimal scheduling. Both the calculation time and accuracy by applying the improved PSO are satisfactory.