Optimal operation of single and a cascade hydro-electricity reservoirs systems were found using genetic algorithm and excel optimization solver and the results were comparatively analyzed. The objective function was to minimize the difference between actual and installed generation capacity of plants. The state transformation equation (the equation of water balance), the minimum and maximum stage and turbine releases were taken as constraints. A random sequence of ten days has been chosen to run the models. The results showed that the release policy of genetic algorithm was better than that of excel optimization solver in two ways: greater electricity generation and convenience of the operation. The impact of population size, number of trials (runs) and number of generations (iterations) on the optimal solution and computing time in genetic algorithm modeling were presented quantitatively. © 2011 Knowledgia Review, Malaysia.