Abstract: In view of the unequal data sampling time interval, the great difficulty of date sampling and a small amount of data in failure prediction of missile equipment, a prediction method based on data fusion and amendatory metabolism unequal interval grey model (AMUGM(1,1)) is proposed according to data fusion technology and grey prediction theory. The unequal interval grey model (UGM(1,1)) is established by improved initial value choice and background value constitution, and is optimized by residual error correction and metabolism．Then a membership grade weighted method is put forward based on weighted idea so as to determine the membership values of the models. Finally, a failure prediction model of special individual is built on the basis of membership values and AMUGM(1,1) model. The simulation result shows that the proposed method is effective.

Key words: ordnance science and technology, failure prediction, grey model, unequal interval grey model, data fusion