The GM(1, 1, ??^??) model with time power is a generalization of the grey GM(1, 1) model. Based on the grey GM(1, 1) model and the equidistance GM(1, 1, ??^??) model with time power, the non-equidistance GM(1, 1, ??^??) model with time power is proposed. The relationship of the model’s curve, power’s exponent and development coefficient is discussed, and the parameter space of non-equidistance GM(1, 1, ??^??) model with time power is studied. The average relative error
is seen as a function of power’s exponent. The numeric area of power’s exponent can been got according to the shape of raw data. The particle swarm optimization(PSO) algorithm is used to solve the power’s exponent. The practical application illustrates the effectiveness of the proposed model.