Abstract: Ammunition supply is the lifeblood of combat troops, active, accurate and rapid implementation of ammunition delivery has become an important growth point in modern warfare, fast and accurate driving route choice and time prediction of ammunition delivery are crucial. Through analyzing the traditional methods of ammunition delivery, using satellite images processing technology, the article establishes the driving route choice and time prediction algorithms of ammunition delivery and application policy model. The traffic flow of the route is the least, it is the best. The speed of traffic flow is used as the speed of delivery vehicles to predict arrival time of the ammunition delivery. In order to improve the reliability and accuracy of traffic flow prediction algorithms, statistical data mining is used to optimize the algorithm parameters, and thus to reduce the algorithms time complexity by use of parallel computing, to achieve optimal route selection of ammunition delivery and reliable, accurate and fast prediction of arrival time.

Key words: ammunition delivery, time prediction, data mining, satellite images