This paper characterizes the operation dispatches of giant hydropower station cluster, whereby to propose variable-step size and chaotic annealing algorithms for optimization based on the objective function nonlinearity of short-term dispatch and the electro-hydraulic connection complexity of long-term dispatch. An example taken for computation proved that the proposed algorithm is highly reasonable and excellent. Short-term optimization dispatch system turns out that the variable-step-size algorithm, as a new approach for dispatching the hydropower station cluster, can reduce the coupling constraints on the dispatch when the electro-hydraulic connection is complex. The efficiency of solution is improved by the fixed step size variable in a simple process. It is feasible to work out the coupling constraints for short-term dispatch by introducing the mutative scale strategy. For the long-term optimization dispatch, it follows that the proposed algorithm can guarantee the global convergence of results and has diversified the populations. Adjustable crossover and mutation operators improve the adaptability of algorithm. The results from simulation reveal that the dispatch optimized by the algorithm proposed in this paper features less time-consuming, increased power generation and the feasible optimization solution.