Reservoir operational rule curves are equations that can balance reservoir system parameters in each period by considering previous experiences of the system. These equations include variables such as inflow, storage capacity, and released water from the reservoir that are commonly related to each other with some constant coefficients in a predefined linear and nonlinear pattern. In the recent years, evolutionary algorithms in general and genetic algorithm (GA) in particular have been applied to develop optimal operational rule curves with predefined pattern. Genetic programming (GP) is an evolutionary algorithm based on GA, which is capable to calculate optimal operational rule curves without using predefined operational patterns. In this paper, at first, first- and second-order rule curves which depend on the inflow and storage volume of Karaj and Bazoft reservoirs have been extracted for the meeting downstream water demands and hydropower energy generation objectives, respectively. In following, operational rule curves by the GP, without any predefined mathematical pattern have been extracted and compared with the GA results. The GP rules improved the objective functions of the first- and second-order rule curves by the GA in supplying downstream demand 12.39 and 7.29 percent, respectively.. Similarly, the hydropower energy generation has been improved by 51.62 and 47.76 percent in the first- and second-order rule curves, respectively.