In order to establish a high-precision nonlinear forecasting model, the paper presents a new global optimization technique for parameters optimization in nonlinear forecasting model based on the minimization of mean absolute percentage error (MAPE). By implementation of an optimization technique based on the successive use of a genetic algorithm and of a sequential quadratic programming (SQP) method: the genetic algorithm is used to perform a preliminary search in the solution space for locating the neighborhood of the solution, then, parameters optimization of nonlinear forecasting model is formulated as the minimax optimization problem by a transition, the SQP method is employed to solve minimax optimization problem for refining the best solution provided by the genetic algorithm, thus achieving the final optimal model and minimum MAPE. Three realworld examples are used to examine the simulation accuracy of the proposed method, the results show that the proposed method can improve the simulation precision compared to the existing algorithm.