AbstractIn view of the consumption prediction of aviation guided ammunition, consumption point prediction and interval prediction models are established respectively.According to the characteristics of small sample, nonlinearity and strong randomness, the support vector regression model is adopted to predict the consumption data, and the optimal parameters are found by particle swarm optimization algorithm. Combining with the point prediction error data, the error uncertainty analysis is conducted by kernel density estimation to determine the error probability density curve, and determine the best confidence interval for a given confidence using highestdensity regions based on kernel density estimation.The results show that the proposed method can provide more accurate prediction results and uncertainty change interval for the use consumption data, and provide a reference for the subsequent use arrangement of aviation guided ammunition.