To solve the complex non-linear problem of material consumption in battlefield, a new self-adaptive forecast model of material consumption in battlefield is established in this paper, based on the analysis of traditional forecasting method, and the self–adaptive mechanism based on Kalman filter is discussed in detail, which is tested by an instance. The result shows that the model can achieve a relatively accurate prediction by overcoming the impact of uncertainties in battlefield, and the error between predictive value and actual value can be controlled within a certain threshold.