Abstract: As the background value is an important factor affecting the precision of grey system model, this paper makes a research on the background value in non-equidistant GM (1,1) model.Based on index characteristic of grey model and the definition of integral, we use the discrete function with non-homogeneous exponential law to fit the accumulated sequence to propose a new method to optimize the background value in non-equidistant GM (1,1) model.The example indicates the precision of simulation and forecast by this optimized background value is obviously higher.This new formula of background value is motonly suitable in non-equidistant GM (1,1) model, but also in equidistant GM (1,1) model, and has high precision and wide applicability.

Key words: non-equidistant, GM (1,1) model, background value, optimize