Abstract: Complicated equipment development cost forecasting of the cross section data is studied, the GM(0, N) model is relatively stable when the sample is small. The data is sorted according to the similarity degree proposed, then GM(0,N) model is established, the more similar to the prediction the object sample data, the more sensitive to GM(0,N) model, the greater weight GM(0,N) model determine from the angle of sensitivity. Due to the similar equipment often produce similar development costs, making full use of the similar sample data to prediction object is helpful to improve the forecasting accuracy. The practical example is given to illustrate the practicability and validity of this model.

Key words: missile development cost, grey model, cross-sectional data, similarity degree